SE539283C2 - Method and system for assessing the trip performance of a driver - Google Patents

Abstract

77 Abstract Method for automatically assessing performance of a driver (110) of a vehicle (100) for a particular trip, wherein current driving data sets, comprising basic driving data are repeat- edly read from the vehicle, which method comprises the steps a) collecting previous-trip driving data sets, comprising instantaneous vehicle energyconsumption and instantaneous vehicle velocity, for different previous trips, differentdrivers and different vehicles; b) for a plurality of said previous-trip data sets, calculating a respective relative instan-taneous vehicle energy consumption value; c) calculating a characteristic vehicle relative energy consumption function; and d) calculating a value of a trip performance parameter based upon a weighted averagevalue of the respective relative instantaneous energy consumptions for previous-tripdata sets that correspond to each of said current-trip data sets, which weighting is performed using said cha racteristic vehicle relative energy consumption function. The invention also relates to a system. Application text 2015-12-15 150120SE

Description

Method and system for assessing the trip performance of a driver The present invention relates to a method for assessing the trip performance of a driver. lnparticular, the invention relates to such assessment in relation to a driver of a vehicle, suchas a motor vehicle. ln some aspects, the invention also relates to such assessment in relationto a driver of a non-motorized vehicle, such as a bike. Furthermore, the invention relates to a system.

Today, an abundance of data is available electronically during and after driving of variousvehicles, such as trip computers providing information about a current trip performed usinga vehicle; standardized digital interfaces, such as a CAN-bus based interface, arranged invehicles and arranged to provide vehicle- and driving-related data to hardware appliancespluggable into the vehicle using such interfaces; and data available from standalone porta-ble equipment, such as smartphones and GPS equipment, arranged in the vehicle duringuse. Such data is today used fortraffic information purposes, by wirelessly collecting currentdriving data for many vehicles, such as using the internet, and calculating expected travel- ling times, performing route planning and so forth.

At the same time, there is an increasing need, for reasons of environmental concern, econ-omy, risk management, etc., of measuring the driving performance of individual vehicledrivers and groups of drivers. For instance, by measuring fuel consumption, it may be pos-sible to determine how environmentally friendly the driving style of a particular driver is. lnthe extension, such information may be used to, for instance, keep track on the total envi-ronmental impact of a fleet of vehicles. Also, such information can be used for feedbackpurposes, in order to improve performance over time for individual drivers as well as on an aggregate level.

However, since different vehicles have typical fuel consumption profiles, and since identicalvehicles can perform very differently under different conditions in terms of load, traffic sit-uation, road conditions, and so forth, only using fuel consumption is a blunt measure. ln addition to this, reliable fuel consumption data is not readily available for many types of vehicles. For non-motorized vehicles, such as bikes, fuel consumption is not relevant at all as a measurement value for a particular trip with such vehicle.

Hence, there is a need for a way to more accurately measure driving performance of vehicledrivers, in particular for individual trips, which may be used both for comparing relativedriving performance from various perspectives between different trips by the same driver/vehicle combination as well as across several drivers/vehicles.

The present invention solves these problems.

Hence, the invention relates to a method for automatically assessing performance of adriver of a current vehicle for a particular current trip, wherein updated current-trip drivingdata sets are repeatedly read from the vehicle, which current-trip data sets each comprisesdata from at least a predetermined set of basic driving data parameters, wherein new suchcurrent-trip data sets are read from the vehicle at consecutive observation time points sep-arated by at the most a predetermined observation time period, which method is charac-terised in that the method comprises the steps of a) collecting previous-trip driving datasets, observed at a plurality of different observation time points, for a plurality of differentprevious trips made by a plurality of different drivers and a plurality of different vehicles,which previous-trip data sets each comprises parameter values for at least a certain prede-termined set of qualified driving data parameters in turn comprising the said basic param-eter set and in particular instantaneous vehicle energy consumption and instantaneous ve-hicle velocity; b) for a plurality of said previous-trip data sets, calculating a respective rela-tive instantaneous vehicle energy consumption value, which relative energy consumptionis relative to a total energy consumption for a respective trip during which the previous-tripdata set in question was observed; c) calculating a characteristic vehicle relative energy con-sumption function regarding the value of said relative instantaneous vehicle energy con-sumption for different instantaneous vehicle velocity parameter values; and d) calculatinga value of a trip performance parameter based upon a weighted average value of the re-spective relative instantaneous energy consumptions for previous-trip data sets that corre- spond to each of said current-trip data sets based upon a similarity or conformance measure regarding the respective values of said basic parameters, which weighting is performed us- ing said characteristic vehicle relative energy consumption function.

The invention further relates to a system for automatically assessing performance of adriver of a current vehicle for a particular current trip, which system is arranged to repeat-edly read updated current-trip driving data sets from the vehicle, which current-trip datasets each comprises data from at least a predetermined set of basic driving data parame-ters, wherein the system is arranged to read new such current-trip data sets from the vehi-cle at consecutive observation time points separated by at the most a predetermined ob-servation time period, which system is characterised in that the system comprises a server,arranged to collect previous-trip driving data sets, observed at a plurality of different obser-vation time points, for a plurality of different previous trips made by a plurality of differentdrivers and a plurality of different vehicles, which previous-trip data sets each comprisesparameter values for at least a certain predetermined set of qualified driving data parame-ters in turn comprising the said basic parameter set and in particular instantaneous vehicleenergy consumption and instantaneous vehicle velocity, in that the system is arranged to,for a plurality of said previous-trip data sets, calculate a respective relative instantaneousvehicle energy consumption value, which relative energy consumption is relative to a totalenergy consumption for a respective trip during which the previous-trip data set in questionwas observed, in that the system is arranged to calculate a characteristic vehicle relativeenergy consumption function regarding the value of said relative instantaneous vehicle en-ergy consumption for different instantaneous vehicle velocity parameter values, and in thatthe system is arranged to calculate a value of a trip performance parameter based upon aweighted average value of the respective relative instantaneous energy consumptions forprevious-trip data sets that correspond to each of said current-trip data sets based upon asimilarity or conformance measure regarding the respective values of said basic parame-ters, which weighting is performed using said characteristic vehicle relative energy con- sumption function. ln the following, the invention will be described in detail, with reference to exemplifying embodiments of the invention and to the enclosed drawings, wherein: Figures 1a-1d are respective Simplified views of a vehicle showing respective parts of a sys-tem according to four different embodiments of the present invention, which systems arearranged to perform a method according to the present invention; Figure 2 is an overview illustration of a system according to the present invention, arrangedto perform a method according to the present invention; Figure 3 is a flowchart illustrating a method according to the present invention; Figure 4 is a flowchart also illustrating a method according to the invention; Figure 5 illustrates a measurement scheme according to the present invention; Figures 6A and 6B illustrate a mapping of a particular vehicle to a particular vehicle class;Figures 7-12 are respective flowcharts illustrating methods according to the present inven-tion; Figures 13a and 13b show respective characteristic instantaneous relative energy consump-tion curves according to the invention; Figures 14 and 15 are respective flowcharts illustrating methods according to the presentinvention; and Figures 16-17 are simplified illustrations of respective exemplifying embodiments of the present invention.

The figures share reference numerals for same or corresponding parts. ln general, the present invention relates to a method and a system for automatically as- sessing performance of a driver of a current vehicle for a particular current trip.

Herein, the term ”performance” relates to a quantifiable, in particular measurable and/orcalculable, quality of a particular trip driven by a particular driver, and in particular in rela-tion to the driving of the driven vehicle as such. A ”performance parameter" or ”perfor-mance measure” is a well-defined parameter the value of which is a measure of the saidquality in some respect ofthe trip in question. Such quality may be in terms of environmen-tal footprint, risk of accidents, driver stress level, vehicle wear or any other quantifiable metric relating to the driven trip.

Furthermore, herein the term ”trip” means a journey performed using a particular vehicleand as controlled by a particular driver ofthe vehicle in question. A trip may be a round tripor a single way trip. When a trip starts and ends may be determined manually by the driver, and/or may be determined automatically based upon location or velocity data, or similar.

A ”current vehicle” means a vehicle which has performed, currently performs or will per-form a ”current trip” in the sense of the present invention, namely a trip for which a perfor-mance measure is calculated or is to be calculated according to one or several of the pre- ferred embodiments described herein.

A ”vehicle” can be a car, a bus, a truck, a motorcycle, or any other motorized vehicle, suchas a gasoline-, diesel- or gas-propelled vehicle, of a vehicle propelled by any other flamma-ble carbohydrate or non-carbohydrate based fuel, comprising an explosion motor; or anelectrically powered vehicle comprising an electrical motor and a battery. lt may also, insome embodiments, be a non-motorized vehicle, such as a bike, a kickbike or rollerskates.The present invention is also applicable to trains, airplanes, helicopters, boats and otherpropelled vehicles travelling on the ground, water or in the air. Such applications are thenimplemented in a respective way which is analogous to what is said herein below in relation to cars, bikes, etc.

A ”driver”, as used herein, is a person controlling the vehicle in question during the trip.Examples comprise the driver ofa car or a bus, as well as a person cycling on a bike. ln someembodiments, a ”driver” may also be an autopilot or other human-assisted machine, oreven a machine arranged to drive the vehicle completely independently. Such machine may be software- and/or hardware implemented, as suitable.

According to the invention, updated current-trip driving data sets are repeatedly read from the vehicle, in particular from a current vehicle.

Such reading can be performed by a piece of hardware equipment which is separate fromthe vehicle, such as a piece of equipment which is physically connected to, and communi-cates via, a hardware interface provided by the vehicle in question, such as a standardizedhardware interface for connecting equipment for vehicle diagnosis or similar. This is illus-trated in figure 1a, in which a piece of hardware 120 is physically connected to a hardwareinterface 101 of a vehicle 100 driven by a driver 110. The piece of hardware 120, which mayfor instance be a convention OBD (OnBoard Diagnostics) reader, communicates with a port-able electronic device 130, such as a mobile phone, a PDA, a laptop computer or similar, using a wireless 121 or wired 122 communication channel.

The portable electronic device 130 may preferably be controlled, by the driver 110, and ispreferably a general-purpose programmable computer device, such as a conventionalsmartphone, with wireless communication capabilities allowing it to communicate wire-lessly with entities located outside of the vehicle 100, such as a base stations 140 of a mobiletelephony network. Preferably, the mobile electronic device 130 comprises a SIM (Sub-scriber Identity I\/|odule) card 131 or corresponding functionality, using which the portableelectronic device 130 identifies itself to such a mobile network. Preferably, the communi-cation between the mobile electronic device 130 and the base station 140 is a digital con-nection, preferably an internet connection, such as using GPRS, 3G, LTE, 4G or 5G. Prefera-bly, the wireless communication 121 between the mobile electronic device 130 and the piece of hardware 120 is a local communication, such as NFC, Bluetooth®, WiFi, or similar.

Figure 1b illustrates an alternative setup, in which a portable electronic device is not re-quired, but wherein the said piece of hardware 120 itself comprises wireless communica-tion functionality 123, such as GPRS, 3G, LTE or WiFi functionality, for communicating withthe base station 140, or a similar entity. Preferably, such communication is based uponidentification using a SIM card 131, or similar functionality, installed in the piece of hard-ware 120, and is preferably a digital communication, preferably an internet communication.However, the communication may also be through a local wireless or wired communication,such as a Bluetooth® or USB interface. ln the latter case, communication between the vehi- cle 100 and the central server 150 will take place intermittently, such as the vehicle 100 loading up data to the central server 150 when parked, such as during charging och refuel- ling.

Figure 1c illustrates yet another alternative setup, in which the piece of hardware 120 is notrequired either. ln this case, the vehicle 100 comprises a piece of hardware 102 having acommunication means 103 arranged to carry out communications with the base station 140as described above, preferably based upon identification using a SIM card 104, or similarfunctionality, installed in the vehicle 100. Preferably, the communication in this case is dig- ital, preferably in the form of a wireless internet connection.

Figure 1a further illustrates a central server 150, which is in contact with the base station140, for instance via a mobile telephony operator, and for instance in addition via a conven-tional internet connection 170. Hence, the central server 150 and the portable electronicdevice 130 are arranged to communicate with each other, at least the portable electronicdevice 130 is arranged to, while arranged in the vehicle 100, provide information to thecentral server 150 using wireless communication. The corresponding is true regarding thepiece of hardware 120 in figure 1b and the piece of hardware 102 in figure 1c, that arecorrespondingly arranged to provide information to the central server 150 using wireless communication from the vehicle 100.

Also in figure 1d, the central server 150 is present, together with the internet connection170 and the base stations 140. However, in figure 1d, there is a server 160 arranged at or inthe vehicle 100 also. Namely, figure 1d illustrates an alternative or supplementary embodi-ment, using a local server 160 arranged at or in the vehicle 100. ln this case, depending onthe practical embodiment as described in further detail below, the said provision of infor-mation to the central server 150 from a corresponding wireless entity 102, 120, 130, asillustrated in figures 1a, 1b, 1c, may or may not take place. Furthermore, a provision of cor-responding information takes place to the local server 160, using a wired or wireless com-munication channel provided locally in the vehicle 100. ln figure 1d, the local server 160 is illustrated as a standalone server, for exemplifying purposes. lt is, however, realized that the local server 160 may be a software component comprised in the mobile electronic de-vice 130. The local server 160 may also be comprised in any of pieces of hardware 102, 120,130 or be in suitable wired or wireless communication with any of pieces of hardware 102,120, 130. I\/|oreover, the local server 160 is preferably arranged to communicate wirelesslywith the central server 150, such as via the portable electronic device 130 and further viabase stations 140 and the internet 170, or using a proprietary, preferably SIM card (or sim- ilar) identification based, communication functionality. ln a particularly preferred embodiment, the local server 160 is integrated in the hardwareof the vehicle 100, in which case devices 120, 130 are not necessary, but the onboard-vehi-cle system functionality is completely self-contained in the vehicle. ln that case, the localsever 160 may provide the driver of the vehicle 100 with trip performance parameter valuefeedback (see below) based upon driving data set data in the local database 161, but may only intermittently communicate with the central server 150. lt is realized that, in case a local server 160 is used, a respective such local server 160 maybe arranged in several different vehicles. Hence, the system according to the present inven- tion may comprise the central server 150 as well as a plurality of local servers 160. ln order for the system to know what vehicle is used, and/or which driver is driving whatvehicle, during a particular observed trip, it is preferred that each driver and/or each vehiclehas an account, such as a user account, on the server 150, which may be registered aheadof time in a way which is conventional as such. lt is further preferred that the user accountis tied to an authenticated session, such as a login session using the portable electronic de-vice 130, the vehicle 100, or in another suitable way, so that the server 150 gains knowledgeabout what user of the system is currently driving the vehicle 100. ln a corresponding man-ner, it is preferred that the vehicle 100 is automatically identified to the server 150, forinstance by the vehicle 100 identifying itself automatically to the server 150 based upon aunique vehicle identity, or by the user selecting the current vehicle from a list of predefined vehicles for that user.

As mentioned above, the present invention relates to a method and a system for assessingthe driving performance ofa driver. This assessment is, in general, performed automaticallyby said system, and in particular by automatic calculations performed mainly in the centralserver 150 and/or in local servers 160. ln the following, all calculations and determinationsare performed automatically unless it is explicitly stated that they involve some type of manual interaction.

As used herein, the term ”driving data set” is a set of parameter data observed during aparticular trip. Preferably, the vehicle performing the trip is arranged to measure corre-sponding parameter data, using suitable hardware and/or software, substantially at the mo-ment in time at which the driving data set is read from the vehicle, so that the driving dataset substantially represents real-time, or at least near real-time, data regarding the trip inquestion, at the respective time or measurement. A ”current-trip driving data set” is such adriving data set read from a current vehicle in relation to a current trip, preferably a current trip which is actually currently being undertaken when the reading in question is performed.

Furthermore according to the invention, the said current-trip driving data sets each com-prises data from at least a predetermined set of basic driving data parameters. Such param-eters constitute measurable data regarding the progression of the trip in question, in thiscase the current trip, in particular such data which is measurable by the respective con- cerned vehicle itself. This will be exemplified below.

The basic parameter set preferably comprises at least 3, more preferably at least 4 differentparameters, and preferably at the most 10, more preferably at the most 7, different param- QTQFS.

Moreover, new such current-trip data sets are read from the vehicle in question at consec-utive observation time points separated by at the most a predetermined observation timeperiod. Preferably, the read updated current-trip data sets are immediately or at least sub-stantially immediately, upon reading, communicated to the central server 150 and/or, de- pending on the embodiment currently performed, to a local server 160, as the case may be, lO and using the above described wired and/or wireless communication links. lt is understoodthat the current-trip data sets are read by the vehicle 100 and made available via inter- face(s) 101, 103, 121, 122 and/or 123.

The current-trip driving data sets are hence collected for, and preferably during, a currenttrip, and are furthermore stored, at least in aggregate form, in the central server 150 and/orin the local server 160 for future reference. For this and other purposes, the central server150 and the local server 160 comprises a respective database 151, 161, for digital storage of read driving data sets.

Hence, preferably, parameter values of said basic parameter set are automatically capturedby the vehicle 100 and communicated to the said portable electronic device 130 arrangedat the vehicle 100, which portable electronic device 130 then communicates, via a wirelesslink 121, 140, said driving data sets to the central server 150. Alternatively, the said param-eter values are communicated, via a wireless link 103, 121, 123 directly from the vehicle 100 to the central server 150.

Figure 2 illustrates three vehicles 100a, 100b, 100c, ofwhich one may be a current vehicle,and of which each may be as illustrated in figures 1a-1d. Each of the vehicles communicates with the central server 150 as described above.

First aspect According to an exemplifying embodiment of the present invention, illustrated in figure 3 in the form of a flow chart, in a first method step the method is started. ln a subsequent method step, previous-trip driving data sets are collected.

As used herein, the term ”previous trip” is a trip which was performed, by a certain vehicle, a ”previous vehicle", at least partly before a current trip is conducted, or at least a trip for which driving data sets are available to the central server 150 before the current-trip driving ll data sets are collected for analysis. ln particular, at least one data set read in relation to and during a previous trip is read before at least one current-trip driving data set is read.

Correspondingly, a ”previous-trip driving data set” is a data set of the type discussed above,read from a vehicle performing a previous trip. lt is realized that a driving data set read bya current vehicle can constitute a previous-trip driving data set for another current vehicle,at a later point in time, or even for the same vehicle, which at that later point in time is a current vehicle.

Such collected previous-trip driving data sets are observed at a plurality of different obser-vation time points, for a plurality of different previous trips made by a plurality of differentdrivers and a plurality of different vehicles. ln particular, it is preferred that said previous-trip driving data sets are observed at at least 1 000, more preferably at least 10 000, evenmore preferably at least 100 000, different observation time points, and/or for at least 100,more preferably at least 1 000, even more preferably at least 10 000 previous trips, and/ormade by at least 5, preferably at least 50, more preferably at least 100, different driversand/or at least 5, preferably at least 50, more preferably at least 100, even more preferably at least 1 000, different vehicles. ln other words, the previous-trip driving data sets collected preferably constitute a large amount of data regarding different trips, drivers and/or vehicles.

The collecting of the previous-trip driving data sets may take place by the vehicle 100a,100b, 100c in question, as described above with respect to the collecting of current-tripdriving data sets, which collecting comprises reading by the vehicle in question, communi-cating to the central server 150 and/or a local (arranged in the vehicle in question) server160, and subsequent storage therein. ln case a local server 160 is used, the stored infor-mation is subsequently provided to the central server 150, using a suitable communicationmethod, such as wirelessly via base stations 140, intermittently or after the previous trip is finished. Hence, the central server 150 will eventually receive and centrally store, in the 12 database 151, a number of previous-trip driving data sets for each previous trip conducted using the method and system according to the present invention.

Each of said previous-trip driving data sets comprises respective parameter values for atleast a certain predetermined set of qualified driving data parameters. Such parametersconstitute, similarly to the above mentioned basic driving data parameters, measurable data regarding the progression ofthe trip in question, in this case the previous trip.

The qualified driving data parameter set comprises the basic parameter set, which basic setis hence a subset of said qualified set. The basic and qualified sets may also be identical. lnparticular, the qualified driving data parameter set comprises instantaneous vehicle energyconsumption. Preferably, however, the basic driving data parameter set, as opposed to thequalified driving data parameter set, does not comprise insta ntaneous vehicle energy con- sumption.

Said instantaneous vehicle energy consumption may be, for instance, instantaneous fuelconsumption or instantaneous power use of a battery used to propel the vehicle in ques-tion. Preferably, the instantaneous vehicle energy consumption is measured and expressedin relation to travelled distance, such as ”litres per km” or ”Wh per km", or the correspond-ing, even though in other embodiments they could also be expressed in relation to time, such as ”litres per hour”.

The above-described collecting of current-trip driving data sets may be performed in paral- lel to the collecting of previous-trip data sets, or afterwards. ln a subsequent method step, which may be performed in parallel to the said collecting ofprevious-trip driving data sets, the collected previous-trip driving data sets are grouped, orclassified, into basic historic groups of such sets. Preferably, each previous-trip data set isclassified into at the most one, preferably exactly one, of said basic historic groups. ln this classification, the said historic groups hence constitute the classes into which the driving 13 data sets are classified. lt is noted that these ”classes” are not the same as the vehicle clas- ses described below.

Preferably, the said grouping is based upon a basic driving data set similarity measure, inother words a comparison measure for comparing driving data sets comprising said basicparameter set, and determining a similarity between such compared data sets. Preferably,the basic driving data set similarity measure is also arranged to compare qualified drivingdata sets one to each other, or even to compare basic driving data sets to qualified drivingdata sets, based upon the values of said basic parameters comprised in such data sets. How-ever, it is preferred that the basic driving data set similarity measure does not take instan-taneous energy consumption data in said driving data sets into consideration for the calcu-lated similarity measure. Further preferably, the same basic driving data set similarity meas- ure is used for all similarity calculations between driving data sets as described herein.

”Similarity”, as used herein with respect to two driving data sets, refers to numerical simi- larity of the respective parameter values of the driving data sets in question.

Hence, after such classification in this method step, a number of basic historic groups willexist, each comprising zero or more previous-trip driving data sets that are sufficiently sim-ilar one to the other according to the said similarity measure. lt is also possible that the saidclassification is performed continuously, so that newly collected previous-trip driving datasets are classified into one of said basic historic groups in connection to their collecting, orintermittently. ln this case, the contents of said basic historic groups will be dynamically updated as time goes by. ln practise, each previous-trip driving data set may not be individually stored in the data-base 151. lnstead, the previous-trip driving data sets for each particular basic historic groupare preferably stored in an aggregated manner in the database 151. This may, for instance,be achieved by the definition of each basic historic group being associated, in the database 151, with corresponding aggregate data calculated based upon the previous-trip driving 14 data sets being mapped to the basic historic group in question. Such aggregate data may comprise, for instance, a group performance parameter value (see below). ln a subsequent method step, for each of said current-trip driving data sets co||ected asdescribed above, the current-trip driving data set in question is mapped to at the most oneparticular ofthe above-described basic historic groups, hence the same set of basic historic groups used for classification ofthe previous-trip driving data sets.

This mapping, of the current-trip driving data sets to said basic historic groups, is basedupon a basic group conformity measure between a driving data set and a basic historicgroup. Even though the said conformity measure is a measure of conformity between a dataset and a group, in other words how closely the data set in question conforms to the groupin question, whereas the above discussed similarity measure is a measure of similarity be-tween one data set and another data set, the conformity measure can be similar to thesimilarity measure, or even analogous in the sense that the requirements for two data setsto be classified into one and the same particular group are the same as the requirementsfor a certain data set to be classified into the group in question. ln particular, it is preferredthat both the conformity measure and the similarity measure is based upon a basic historicgroup definition, so that ”conformity” between a data set and a group means that the dataset falls under the definition of the group, whereas ”similarity” between two data setsmeans that both data sets fall under the definition of one and the same group (regardlesswhich such group). This will make possible a calculationally simple implementation, result- ing in a high performance system.

According to a particularly preferred embodiment, the basic driving data set similaritymeasure is arranged to classify driving data sets into one of a plurality of different prede-termined basic historic groups based upon conformance of the respective basic parametervalues comprised in the driving data set in question to respective allowed parameter valueranges for each of said parameters. ln particular, for each parameter in said basic set ofparameters and for each basic group, a predefined respective parameter value range is de- fined. Then, the basic group is defined in terms of a combination of one such parameter value range for each parameter in the basic set of parameters. lt is noted that not all pa-rameters in said basic set of parameters necessarily have to be used, in other words one orseveral parameters can have very large allowed intervals. Preferably, at least two, moreprefera bly at least three, pa ra meters in the basic set of parameters are associated with con-secutive, non-overlapping intervals, and that different groups are defined by unique com-binations of such mutually non-overlapping intervals of parameter values. lt is preferredthat, for each such parameter, there are at least ten, preferably at least fifty, such non- overlapping intervals. lt is realized that other similarity measures can also be used, such as some type of geometricdistance measure based upon the numerical values of the parameter values in each data set, in this case viewed as a vector of values. ln a manner corresponding to the above described, with respect to the similarity measure,the basic group conformity measure is arranged to classify driving data sets into one of saidplurality of different predetermined basic historic groups based upon conformance of therespective basic pa ra meter values comprised in the driving data set in question to respec-tive allowed parameter value ranges for each of said parameters. Preferably, the intervalsused to define the basic conformity measure are the same intervals as used to define the said basic similarity measure. lt is preferred that the said conformity measure is based upon the numerical parametervalues of the current-trip driving data set in question and upon a definition of the basic historic group in question. ln subsequent method steps, a first energy consumption-based trip performance pa rameter value is calculated for the current trip.

Preferably, the first energy consumption-based trip performance parameter is calculatedbased upon a respective energy consumption-based group performance parameter value for each of the basic historic groups to which at least one current-trip driving data set was 16 mapped as described above. Hence, in a first of said subsequent method steps, or before-hand, such a group performance parameter value is calculated for at least each such mapped basic historic groups. lt is realized that the above-described grouping of previous-trip driving data sets into basichistoric groups; mapping of current-trip data sets to said groups; and/or calculation of saidgroup performance parameter values can be performed on the fly, continuously as newdata becomes available. lt is preferred that the said group performance parameter valuealways takes into consideration all previous-trip driving data sets available to the entity per-forming the said calculation, however not that the current-trip driving data set values areused as previous-trip driving data sets before the current trip is finished. Upon finishing, thecurrent trip may become a previous trip for a subsequent current trip, ofthe same or an- other vehicle and/or user.

For each mapped basic historic group, the said group performance parameter value is cal-culated based upon the respective instantaneous energy consumption value in the respec-tive previous-trip driving data sets classified into, and therefore comprised in, the basic his-toric group in question. lt is noted that the previous-trip driving data sets, each comprisingsaid qualified parameter set, comprise such a respective instantaneous energy consumption value.

Using such a method and such a system, driving-related data from many different previoustrips, partaken using many different vehicles and by many different drivers, can be used toautomatically assess the driving performance of a current trip in a way which does not re-quire any detailed assumptions of the conditions under which the data was collected. lnparticular, it is possible to obtain a surprisingly accurate view on the performance of thedriver, in terms of driving energy consumption, under very shifting external conditions, interms of for instance vehicle load, road and weather conditions. Furthermore, it is possibleto compare the obtained energy consumption performance across different drivers and also across different types of vehicles. 17 Furthermore, it is possible to achieve these advantages even without any a priori infor-mation regarding neither geography, nor road or traffic conditions. Hence, no expensivemeasurements are necessary; instead, all drivers using the system create a common set ofdata as they use the system, irrespectively of the details describing the external environ- ment in which they perform when doing so.

All these advantages are achievable automatically, without any manual intervention and simply by using the system. This will be explained in further detail below. ln the embodiment described above, this is possible by the use ofthe above-explained basichistoric groups, that are used to disconnect the identity of the previous trips from the pre-vious-trip driving data sets observed during the previous trip in question, and that allowusing the information content in the said data sets irrespectively of the other properties ofeach previous trip. ln particular, this can be achieved in various different aspects, using var-ious detailed techniques as will be described in the following. ln some ofthese aspects, it is not vitally important to use basic historic groups, as will become clear. ln particular, the present inventor has discovered that, by fragmenting a large number ofprevious trips into small segments, where each segment is so small so that essentially noqualitative information can be derived about the driving from such individual segment, andthen mapping the segments of a current trip to such historically collected segments, very accurate information can be derived, in the aggregate, about the current trip. lt is preferred that the calculations described herein are performed using and based uponall, or substantially all, available data from all trips performed by all vehicles that are con-nected to the system. ln this case, the calculations must be performed by an entity havingaccess, in aggregate or detailed form, to all such data. lt is preferred that this entity is thecentral server 150, which then receives previous-trip and current-trip driving data sets fromall connected vehicles, either continuously or intermittently, or differently for different con- nected vehicles, and performs the calculations described above. lt is also possible that the 18 local server 160 receives the previous-trip driving data sets, or corresponding data in aggre-gate format, such as basic historic group definitions together with calculated, preferablyupdated, respective group performance parameter values, from the central server 150, andthen performs the actual calculation ofthe said first performance parameterfor the currenttrip performed by the current vehicle in which the local server 160 is arranged. ln this case,updated data can be provided to the local server 160 only intermittently, such as beforeeach current trip, once a day or even once a week, allowing the advantages ofthe presentinvention to be achieved even when there is no reliable internet connection available to thecurrent vehicle, or when driving abroad without any roaming-based wireless internet con-nection. The said data may even be provided to the local server 160 only once, such as in connection to an installation of a piece of local server 160 software.

According to a preferred embodiment, the said first group performance parameter is a rel-ative instantaneous energy consumption value for the respective previous-trip driving datasets in the basic historic group in question. ln particular, it is preferred that this relativevalue is calculated in relation to a respective total energy consumption for the completeprevious trip during which the previous-trip driving data set in question was observed, suchas a total petrol, diesel, gas or electricity consumption or a total average petrol, diesel, gasor electricity consumption per km of the whole previous trip. ln this case, instantaneousenergy consumption data, as well as total energy consumption data for a whole trip, areavailable from the previous vehicle in question, for reading and submission to the central server 150.

Figure 4 illustrates how this relative instantaneous energy consumption based group per- formance parameter can be calculated. ln a first method step, previous-trip driving data sets are collected for a particular previoustrip. Once all previous-trip driving data sets have been collected for the previous trip inquestion, the total energy consumption for the entire previous trip is also available from the previous vehicle, for reading and collecting as described above. 19 Then, for each such collected previous-trip data set for said previous trip, a correspondingbasic group is identified, and the relative instantaneous energy consumption value for thedata set in question is calculated, for instance by dividing the instantaneous energy con-sumption value for the data set in question by the total energy consumption for the entire previous trip in question.

Then, the group performance parameter for the said basic historic group is updated usingthe corresponding previous-trip driving data set basic parameter values. This may, for in-stance, be performed so that the group performance parameter value is always an averagevalue of the respective calculated relative instantaneous energy consumption values foreach previous trip data set that has, up to that point in time, been allotted to the basic groupin question. For instance, this may be performed by, for each basic historic group, keepingtrack of the number of previous-trip driving data sets that have been allotted to the groupin question, and to perform a suitable weighted average calculation when updating the said relative energy consumption value for the group. ln parallel to the constantly ongoing collecting and evaluating of previous trips and theirrespective data sets, as well as relative instantaneous energy consumption values for vari-ous allotted basic historic groups, current-trip driving data sets are collected as describedabove. Once the first performance parameter value is to be calculated, the collected cur-rent-trip driving data sets are mapped to respective basic groups, as also described above,and the first performance parameter is then calculated based upon the calculated groupperformance parameters. This may, for instance, take place by calculating an average valueof the respective group performance values for all mapped groups for the current trip. Suchaverage value may be a simple geometrical average, or, preferably, a weighted average inwhich the respective group performance values of more frequently updated (by previoustrip data sets being allotted thereto) basic groups are given more weight than respectivegroup performance values of less frequently updated basic groups. ln case not all current-trip driving data sets correspond to a respective existing basic historic group, the averagingfunction may ignore those current-trip driving data sets for the purpose of calculating the first performance parameter value.

Hence, in this preferred example, the first trip performance parameter is calculated basedupon an average value ofthe respective relative instantaneous energy consumption valuesfor the respective basic historic groups to which the respective current trip driving data setsof the current trip have been mapped. The relative instantaneous energy consumptionvalue, in turn, for each basic historic group in question, is calculated based upon an averageof the respective relative instantaneous energy consumption values for each previous-tripdata set allotted to the basic historic group in question, as measured in relation to a corre- sponding previous trip during which the data set in question was observed.

The method according to the present invention can be made completely automatic, collect-ing driving data sets for all trips performed by all vehicles connected to the system accordingto the present invention. However, in order to increase data quality and decrease adverseeffects due to data noise, it is preferred that only a subset of said vehicles are marked astrusted by the system. ln this case, the above described group performance parameter val-ues are calculated so that they are not affected by instantaneous energy consumptions re-ported by vehicles not marked as trusted. For such non-trusted vehicles, driving data setsmay be still be collected, but such collected driving data sets do not affect the group per-formance parameters for the various basic historic groups to which the driving data sets areallotted as described above. Also, non-trusted vehicles may constitute current vehicles, andcurrent-trip performance parameter values may be calculated for such non-trusted vehi-cles. Which vehicles that are to be trusted may, for instance, be manually selected basedupon knowledge about data quality available from particular vehicles and possibly also forparticular drivers; automatically selected based upon data availability for a predetermined minimum data type set for vehicles; or in any other way.

Preferably, the predetermined observation time period mentioned above is relative short,so that each current trip would typically result in a large plurality of different current tripdriving data sets. Preferably, the predetermined observation time period is at the most 10seconds, preferably at the most 5 seconds, more preferably at the most 2 seconds, even more preferably between 0.2 and 2 seconds, most preferably about 1 second. Using such 21 short time intervals strikes a good balance between collecting as much relevant data aspossible while not giving rise to unnecessarily large amounts of data to communicate, store and process. ln particular, it is preferred that the current trip driving data sets are read at regular timeintervals, so that the time period between two consecutive readings is substantially the same for all pairs of such consecutive readings.

The reading as such may be an instantaneous readout or an average value read across anaveraging certain time period, which is preferably at the most 5 seconds of length, prefera- bly at the most 2 seconds of length. lt is noted, firstly, that the collecting ofthe data sets may be performed more intermittently,and also at a certain delay, as long as the reading of driving data sets are performed regu-larly. Secondly, the data set reading frequency may or may not be different from the sam-pling time period length for each read data set. For instance, if readings take place every 1seconds, each reading may relate to vehicle parameter data covering a respective historicaltime period of 5 seconds running up to the currently read second. Such prolonged samplingtime period may be achieved by the vehicle 100 itself, but preferably software and/or hard-ware implemented logic performing such prolonged sampling is comprised in either of de-vices 120, 130, or alternatively in the central server 150 or even in the local server 160. lnthe latter case, the sampling may in practise take place by either device 120, 130, 150, 160receiving repeated instantaneous readings, and performing prolonged sampling readings artificially by performing calculations based upon such repeated readings.Correspondingly, the same is preferred as concerns previous-trip driving data sets. Prefera- bly, current-trip and previous-trip driving data sets are read in substantially the same way, using the same observation time periods.

Second aspect 22 According to one aspect of the present invention, the above discussed basic parameter setcomprises instantaneous vehicle velocity, instantaneous vehicle engine rotation speed, in-stantaneous vehicle velocity change as well as instantaneous vehicle engine rotation speedchange. lt is then preferred that all these parameter values, for each data set in question,are used for calculating said basic group conformity measure, preferably as well as said basicsimilarity measure. The said instantaneous vehicle velocity and instantaneous vehicle en-gine rotation speed are preferably measured on the engine of the vehicle, and preferablyby the vehicle itself, as opposed to being measured using a system which is not connectedto the engine of the vehicle, such as using a GPS-enabled measurement device or similar.GPS-based measurements are hence preferably not used in this context, but only for pro-ducing the below-described extended driving data sets. Said velocity change and enginespeed change are also, in a similar manner, either measured on the engine or calculatedbased upon said instantaneous velocity and engine speed values measured on the engine.Herein, the expression ”measured on the engine” encompasses also other measurementsperformed directly on the vehicle hardware as such, for instance measurements performed on wheels or wheel axes of the vehicle.

Herein, for wheeled vehicles using an explosion motor for propelling of the vehicles, therelationship between instantaneous vehicle velocity and instantaneous engine rotationspeed, and in applicable cases the respective absolute values ofthese two parameters andalso their respective changes over time, have proven to be very useful to consider for thepresent purposes. However, for electrically propelled wheeled vehicles, it is, as an alterna-tive, possible to instead of insta ntaneous vehicle speed use insta ntaneous energy consump-tion, such as instantaneous electrical power usage of the electrical motor propelling thevehicle, as provided by a battery in the vehicle. For the same electrical vehicles, instantane-ous motor load should then be used instead of instantaneous motor rotation speed. Corre-spondingly, and as applicable, instantaneous energy consumption change and instantane-ous motor load change should then be used instead of instantaneous vehicle speed changeand instantaneous engine rotation speed change. Of course, for some vehicle types com-prising both an explosion engine and an electrical motor, both these options can be used at the same time. ln particular, and especially for electrically propelled vehicles, it is preferred 23 that the basic parameter set comprises, in addition to said instantaneous energy consump-tion, instantaneous energy consumption change, instantaneous motor load and instantane-ous motor load change, also use instantaneous vehicle velocity, and preferably also instan-taneous vehicle velocity change. The latter two can be as described above. lt is noted that, for all these values, they are preferably measured on the vehicle as described above.

Herein, whenever instantaneous vehicle speed and instantaneous engine rotation speed,or the corresponding change measures, are used for some purpose, it is in general the casethat, instead or in addition to these values, as applicable, instantaneous motor load andinstantaneous energy consumption may be used correspondingly. This applies both to thepresent aspect, the below described class-defining parameters and elsewhere in this de- scription.

The present inventor has discovered that it is sufficient to use these four basic parametersin order to achieve very reliable data in terms of the said current-drive performance param-eter. ln particular, this is true in case very many previous-trip driving data sets are used for many different vehicles and/or many different drivers, as quantified above.

Since the current instantaneous vehicle velocity, as well as the current instantaneous enginerotation speed, are typically available for readout from the vehicle, they can be readily col-lected. The vehicle velocity change and the engine speed change can be readily calculatedbased upon the said read values, such as by a software and/or hardware implemented logic in any of devices 120, 130, 150 or 160.

Even more preferably, no other data values, apart from the said data values regarding in-stantaneous velocity, instantaneous engine rotation speed, as well as instantaneous veloc-ity change and instantaneous engine rotation speed change, are used by said basic groupconformity measure, and preferably the corresponding is also true for said basic similaritymeasure. This provides for a particularly simple data collecting and performance parameter calculation process, which still is able to provide a high quality output. 24 According to a preferred embodiment, the said instantaneous velocity change is measuredover a certain velocity change time period, so that the velocity change is measured as a velocity difference between two points in time separated by said time period.

Correspondingly, the said instantaneous engine rotation speed change is preferably meas-ured over a certain speed change time period, and hence measured as a difference in in- stantaneous engine speed between two points in time separated by said time period. ln particular, it is preferred that the length of the velocity change time period is differentfrom the length of the speed change time period, while, for each previous-trip driving dataset the corresponding velocity change time period and the corresponding speed changetime period are overlapping. Preferably, the point in time at which the instantaneous vehi-cle velocity is measured and the point in time at which the instantaneous engine speed ismeasured are both, independently ofeach other, contained in both said overlapping changetime periods, and preferably measured at the same or substantially the same time. Suchoverlapping and containing guarantees that the measured parameters of each driving dataset are related to one and the same driving situation, which is important even for very fre- quently measured driving data sets.

I\/|oreover, the said instantaneous vehicle velocity change time period and said instantane-ous engine speed change time period are of different lengths. Namely, in many applicationsit is necessary to fine-tune each of said change time periods to capture relevant data re-garding the trip in question, and in general the optimal change time periods will not be thesame for different parameters. For instance, it is in general preferred that the engine veloc-ity change time period is shorter, such as at least twice as short, as the vehicle velocitychange time period. This is not only due to the fact that engine velocity can change quickerduring driving than vehicle velocity, but also since the use of such shorter change time pe-riod results in the capability of the system to more accurately capture certain driver behav- iour in certain situations while driving.

Even in case the velocity time period and the speed time periods have different lengths, itis preferred that, for each observation time point and hence for each driving data set, they share the same starting time point, or, alternatively the same ending time point.

Furthermore, it is preferred that at least one of said velocity change time period and saidspeed change time periods have a length which is longer than the above discussed prede- termined time period (the time period between each consecutive observation time point).

Both the velocity change time period and the engine speed change time period may havean end point at the corresponding observation time period, and hence correspond to ameasurement conducted in the historical time period running up to the measurement timeof the instantaneous vehicle velocity and the engine speed. However, it is preferred thatthe vehicle velocity change time period runs from the observation time point of said instan-taneous vehicle velocity and forwards, and/or that that the engine speed change time pe-riod runs from the observation time point of said instantaneous engine rotation speed for-wa rds. This results in that each driving data set comprises information regarding the currentsituation in terms of instantaneous velocity and engine speed, as well as how that situationis changed during the coming time period. The present inventor has discovered that this provides very useful performance parameter values for the purposes discussed below. ln case the observation time period is up to about 2 seconds, it is particularly preferred thatthe vehicle velocity change time period for each observation time point starts at the instan-taneous vehicle velocity observation time point and runs forwards, between 3 and 10 sec-onds, and that the engine speed change time period for each observation time point startsat the instantaneous engine speed observation time point and runs forwards, between 1 and 5 seconds.

Figure 5 illustrates an exemplifying measurement scheme for use with a vehicle and thesystem according to the present invention. Along the time axis, a number of consecutiveobservation time points OT1, [...], OT5 are shown, each separated by an observation time period OTP of fixed length. 26 For each observation time point, the following readings are made from the vehicle: 0 lnstantaneous readings (IR1, [...], |R5]) regarding instantaneous vehicle velocity, in-stantaneous engine speed, and any other instantaneously measured values. 0 Engine speed change (ESC1, [...], ESC4). This parameter is measured forward, across atime period which is identical to the observation time period OTP. Hence, the enginespeed change value for observation time point OT1 will not be available until obser-vation time point OT2, and can then be collected as described above. 0 Vehicle velocity speed change (VVC1, [...], VVC4). This pa ra meter is measured forward,across a time period which is longer than the observation time period OTP. Hence, asseen in figure 5, the vehicle velocity speed change value for observation time pointOT1 will not be available until sometime between observation time point OT3 and observation time point OT4, and can then be collected as described above.

Preferably, each driving data set is not collected and used to update a respective basic his-toric group, as described above, until all parameter values are available for the driving data set in question.

Apart from instantaneous vehicle velocity and engine speed, and vehicle velocity- and en-gine speed change, other parameters may also be measured, and may also belong to saidbasic parameter set. Such parameters comprise instantaneous break (either binary on/offor a break force value); instantaneous altitude; altitude change; instantaneous GPS location,altitude or heading, and/or GPS altitude or heading change, and/or GPS altitude accelera-tion; GPS-coordinate based vehicle velocity and/or acceleration; instantaneous engine oiltemperature; gear number used; vehicle blinkers activated; outdoors temperature; statusof cruise control systems; and/or any other data which is available either from the vehicle100 itself or from the mobile device 130 and sensors arranged therein, preferably data that in some respect quantifies the position, behaviour and/or internal state of the vehicle. 27 ln particular, it is preferred that break information, at least in the form of a binary signal(break activated/not activated) is part of said basic parameter set, and is hence also read from the vehicle at each observation time point.

Correspondingly, the said basic similarity and/or conformance measures can take into con-sideration additional parameters of the exemplified types, using the corresponding ap-proach as described above. For instance, in case break information is used in said measure,the binary break value (on/off) may be one ofthe defining parameters of said basic historicgroups, and two driving data sets may be allotted to different basic historic groups in case the driving data sets are identical apart from a difference in break parameter value.

As described above, there is a communication from the current vehicle 100 to the centralserver 150, comprising current-trip driving data sets. ln addition thereto, according to a pre-ferred embodiment, the current-trip performance parameter value is calculated, preferablyas described above by the central server 150, and thereafter communicated, via the abovedescribed wireless link, from the central server 150 to the current vehicle 100, such as tothe portable electronic device 130 arranged at the current vehicle 100, and presented to the current driver. This presentation will be exemplified below.

Such calculation, together with possible communication and presentation to the currentdriver, may be performed in connection to a current trip being completed. However, ac-cording to a preferred embodiment a value of the above described current-trip perfor-mance parameter is calculated repeatedly, preferably at least every 10 minutes, more pref-erably at least every 2 minutes, more preferably at least every 30 seconds, during the cur-rent trip. Then, it is preferred that, for the purposes of calculating the said current-trip per-formance parameter value, the current trip is considered to be that part of the current tripwhich has taken place, and has been collected, up to the moment at which the value ofthecurrent-trip performance parameter is calculated. Hence, the performance parameter valueis calculated as if the collected current-trip driving data sets up to the point of calculationofthe performance parameter value constitute data of the entire, completed current trip. ln this case, it is preferred that the currently calculated such performance parameter value 28 is communicated to the current vehicle and presented to the driver upon each such calcu-lation. This way, the current driver can be provided with regularly updated information re-garding the performance ofthe current trip, which makes it possible for the current driverto adjust his or her driving style in response to such information fed back from the system according to the present invention. ln case the current vehicle lacks an active internet connection, the collected current-tripdriving data sets may be stored locally in the vehicle during the trip, for subsequent uploadto the server 150 once an internet connection is again available. Then, the performanceparameter value can be calculated and provided to the user in connection to this later point in time.

Third aspect ln one aspect of the present invention, illustrated in figure 12, a respective instantaneousrelative vehicle energy consumption value is calculated for a plurality of the previous-tripdata sets as described above. ln particular, throughout the description of this aspect, thisrelative energy consumption is relative to a total energy consumption for a respective trip during which the previous-trip data set in question was observed.

Hence, in a first step, previous-trip and current-trip driving data sets are collected, and thecollected current-trip driving data sets are each mapped to respective previous-trip drivingdata sets in a suitable way, such as using the said basic similarity measure, and/or using basic historic groups and the basic conformance measure, as described above. ln a second step, which may be performed at any time before a third step, and in particularbefore, during or after the said first step, a characteristic relative vehicle energy consump-tion function, regarding the value of said instantaneous relative vehicle energy consumptionfor different instantaneous vehicle velocity parameter values, is calculated. This character- istic function is preferably calculated based upon available previous-trip driving data sets 29 for previous vehicles as explained below. Preferably, there is maximally one such character-istic function for each of the below described vehicle classes, and it is preferred that eachcharacteristic function is updated automatically as new previous-trip driving data sets be-come available, or at least intermittently based upon newly available data. This way, anautomatic compensatory mechanism is accomplished without adding more than limited cal- culation overhead to the method.

The characteristic vehicle relative energy consumption function is preferably not calculatedonly based upon data observed for one vehicle, such as the current vehicle. lnstead, it ispreferably calculated based upon data observed for a plurality of previous vehicles. Thecharacteristic function may be calculated taking into consideration previous-trip drivingdata sets for substantially all, or at least a plurality of, the vehicles in the same vehicle classas the one to which the current vehicle is mapped and to no other vehicles; alternatively itmay be calculated based on previous-trip driving data sets for substantially all, or at least a plurality, of all the previous vehicles regardless of vehicle class. ln said third step, the value of a trip performance parameter, such as the above first orsecond trip performance parameter, is calculated, for instance as described above. ln par-ticular, the performance parameter value is calculated based upon an average value of therespective relative instantaneous energy consumptions for previous-trip driving data setsthat correspond to each of said current-trip driving data sets based upon a similarity or con- formance measure regarding the respective values of said basic parameters. ln case classes are used (see below), the current vehicle is hence first classified into a par-ticular current class of the below described set of classes based upon said class conformitymeasure, and the trip performance parameter value is then calculated based upon only thesaid respective relative instantaneous vehicle energy consumption values for previous-trip data sets in the current collection, corresponding to the current class as defined below.

According to the present aspect of the invention, however, the said average is a weightedaverage wherein the weighting is performed using said characteristic vehicle relative energy consumption function.

Namely, the characteristic vehicle relative energy consumption function describes a char-acteristic relationship between instantaneous velocity and instantaneous relative energyconsumption for previous vehicles. ln other words, for each of a plurality of instantaneousvelocity values or intervals, the characteristic function provides a value of a characteristicor typical relative energy consumption value for the vehicle velocity in question, where eachsuch relative energy consumption value is a relative energy consumption value for previous-trip driving data sets describing the said instantaneous vehicle velocity and in relation to atotal energy consumption during the complete trip during which such a previous-trip driving data set was observed.

When using this characteristic function in order to perform a weighted average calculationwith respect to the relative energy consumptions of each previous-trip driving data set cor-responding to each of the current-trip driving data set, the result is that systematic artefactsrelated to relative energy consumption for different vehicle velocities are automatically cor-rected for, and the reliability of the resulting trip performance value is increased as a result.Examples of possible systematic artefacts comprise systematically high relative energy con-sumption values at low velocities, due to internal engine friction, as well as systematicallyhigh relative energy consumption values at high velocities, due to air friction. However,other artefacts may also occur, such as artefacts only occurring in particular vehicle classes and so forth.

Preferably, the characteristic vehicle relative energy consumption function is normalized,so that its mean value, for all occurring vehicle velocity values or intervals, is 1. This type ofcurve, which is exemplified in figures 13A and 13B, provides for a simple weighted averagecalculation, in which a simple multiplication with the characteristic vehicle relative energy consumption function is often sufficient. 31 As seen in figures 13A and 13B, the average value of the characteristic vehicle relative en-ergy consumption function CHAR, as seen across the whole allowable or used vehicle veloc- ity range, averages to 1, as indicated by the horizontal line "1".

Figure 13A is a continuous function, which may be produced by, for instance, adjusting apolygon function of suitable power to best fit a data set com prising, for said previous-tripdriving data sets, all observed value pairs (instantaneous vehicle velocity; instantaneousrelative energy consumption), and then normalizing the function. Then, for each instanta-neous vehicle velocity, a respective characteristic relative energy consumption is indicated by the characteristic function.

Figure 13B illustrates an alternative way of calculating the characteristic function, in whichthe function CHAR is a step function corresponding to the one illustrated in figure 13A. Thisapproach is particularly advantageous when using the above described interval based basicgroup conformance measure and driving data set similarity measures. Hence, for each of anumber of, preferably non-overlapping and preferably predetermined, vehicle velocity in-tervals (illustrated in figure 13B using vertical lines), the function specifies a respective char-acteristic instantaneous relative energy consumption value. Apart from this difference, the curve illustrated in figure 13B is used in a way which fully corresponds to the curve in 13A. ln particular, it is preferred that the characteristic vehicle relative energy consumption func-tion is calculated based upon an average relative instantaneous vehicle energy consumptionfor several previous-trip data sets having the same vehicle velocity. ln this context, ”thesame velocity" encompasses velocities belonging to the same velocity interval as shown forinstance in figure 13B. Preferably, the same velocity intervals are used for the characteristic curve as those described above in the basic driving data set similarity measure.

Preferably, the characteristic vehicle relative energy consumption function is calculatedbased upon a plurality of basic historic groups of the above defined type, and specificallybased upon a respective value of said relative instantaneous vehicle energy consumption for the previous-trip data sets belonging to the respective basic historic group in question. 32 ln particular, this pertains to each individual vehicle velocity value used for calculating the function, or to each velocity interval covered by the function, as applicable.

The average value is preferably a geometric average. ln case basic historic groups are used,as described above, wherein the said relative energy consumption-based group perfor-mance measure is calculated, the group performance measure can be used, preferably as itis, as the relative energy consumption for calculating the characteristic curve. Furthermoreeach relative energy consumption value in the characteristic curve, or each characteristiccurve, is also preferably calculated as a weighted average, so that more frequently updatedbasic historic groups are given larger weight than less frequently updated basic historic gFOUpS. lt is preferred that the basic parameter set does not comprise a parameter indicating thevehicle type, such as a VIN (Vehicle Identification Number) of the vehicle in question, butthat the only way of characterising the vehicle is using the above described collections and classes.

Fourth aspect According to one aspect ofthe present invention, in order to be able to provide as relevantdata as possible when calculating the said driving performance parameter, at least some,preferably substa ntially all, most preferably all of said previous-trip data sets are classifiedinto a set of collections. ln the above described case in which basic historic groups are used,these collections are used in addition to the basic historic groups, and the previous-tripdriving data sets are hence classified into both a respective basic historic group and a re-spective collection. As will be described in the following, this may take place by each collec-tion comprising its own set of basic historic groups, which sets may then be overlappingbetween different collections, and by each previous-trip data set first being classified into a collection and thereafter into a basic historic group within that collection. 33 lt is preferred that each of said collections only comprises previous-trip driving data sets fora particular class of vehicles, and that all previous-trip data sets ofone and the same vehicleare classified into one and the same collection based upon a basic class conformity measurebetween driving data sets for the vehicle in question and a set of class-defining parameters.ln other words, each vehicle can be characterised based upon driving data sets observed forthat vehicle. ln particular, such driving data sets can be used to determine to what vehicle class that vehicle belongs.

The basic class conformity measure is hence a measure of the conformity of a number ofindividual driving data sets for one and the same vehicle to a particular vehicle class, basedupon the said class-defining parameters for the class in question. After a vehicle has beenassociated with a particular class, each driving data set for the vehicle in question is thenallotted to the same collection, namely the collection corresponding to the vehicle class towhich the vehicle is associated. ln case a vehicle is associated to a particular class at onepoint, and is then reclassified to a different class at a later, second point, the previous-tripdriving data sets collected for that vehicle and already allotted to the collection correspond-ing to the previously associated class can either be reclassified into the collection corre-sponding to the new associated class, or alternatively only newly collected previous-trip driving data sets can be allotted to the new collection.

Each collection may correspond to exactly one class of vehicles, and vice versa.

According to this aspect of the current invention, illustrated in figure 7, before calculatingthe above-described energy consumption-based trip performance parameter, the currentvehicle is classified into a particular one vehicle class, in the following denoted the ”currentclass", of said set of classes, based upon the said basic class conformity measure. Then, thecollection or a collection (the ”current collection") corresponding to the current class isidentified. The current collection preferably comprises all previous-trip driving data sets previously observed for all vehicles currently allotted to the current vehicle class. 34 lt is preferred that each current-trip driving data set is mapped, as described above, to atthe most one particular basic historic group belonging to the current collection, based upon the above described basic group conformity measure.

Then, an energy consumption-based trip performance parameter value is calculated for thecurrent trip. This calculation is preferably performed as described above, and in particularbased upon respective energy consumption-based performance parameter values for pre-vious-trip driving data sets only in the said current collection, as opposed to using all previ-ous-trip driving data sets. Hence, the calculation may be performed based upon a set ofbasic historic groups to which the current-trip driving data sets are mapped, and corre-sponding instantaneous relative energy consumption for such basic historic groups, as de-scribed above, but wherein said set of basic history groups have all been allotted to thecurrent collection. Hence, when mapping the current-trip driving data sets to basic historicgroups as described above, in this case only basic historic groups of the current collection are considered.

Further according to this aspect of the invention, the said basic parameter set further com-prises instantaneous vehicle velocity and instantaneous engine rotation speed. Then, thesaid class-defining parameters comprise, for each class of vehicles, a characteristic enginerotation speed for a particular vehicle velocity. As indicated above, instead of, or in additionto, the said basic parameter set further comprising instantaneous vehicle velocity and in-stantaneous engine rotation speed, it may comprise instantaneous motor load and instan-taneous energy consumption. Then, the said class-defining parameters comprise, for each class of vehicles, a characteristic energy consumption for a particular motor load.

By using instantaneous vehicle velocity and instantaneous engine rotation speed at eachsuch vehicle velocity for defining vehicle classes, and in particular by using only these datafor defining vehicle classes, a vehicle classification yielding surprisingly accurate results interms of driving performance parameter values is achieved. This is in particular the casewhen using a methodology as the one described herein for calculating and processing such parameter values. Furthermore, the invention will produce relevant results even if one and the same vehicle is driven under very different conditions, such as with or without a trailer, on icy or dry roads, with various wind strengths, outdoors temperatures, and so forth.

Figure 6a i||ustrates an example of an empirically or experimentally measured curve C, de-fining a typical or average relationship between vehicle velocity and engine rotation speed,for a particular class of vehicles. The curve C may be determined by, for instance, using alldriving data sets in the collection corresponding to the class in question, for each instanta- neous vehicle velocity calculating an average engine rotation speed.

At a particular point P, a particular instantaneous vehicle velocity VEL corresponds to a par-ticular instantaneous engine rotation speed RS. Hence, P is an example of a class-definingparameterfor that particular class of vehicles. This point P could be calculated as an averageengine rotation speed for all observed driving data sets in the collection in question, that is for vehicles in the class in question, and having the instantaneous vehicle velocity VEL.

A class conformance measure could then, for example, be constructed and used as follows: 1) Construct a curve, corresponding to curve C, but for an individual particular vehiclethe conformance of which is to be determined. The curve is constructed based upondriving data sets observed for that vehicle, such as by taking the average observedengine rotation speed for each observed velocity and then adjusting a polygon func-tion to best fit the achieved pairs of data points for all velocities. 2) For the particular velocity VEL, calculate the distance between the constructed curveand the point P. 3) ln case the distance is smaller than a predetermined largest allowable distance, theconformance measure turns out in the positive, and the vehicle in question is allotted to the class of vehicles in question.

Figure 6b i||ustrates a more complicated, and preferred, embodiment, in which the said class-defining parameters comprise, for each class of vehicles, a respective characteristic 36 engine rotation speed for a plurality of vehicle velocities. ln this exemplifying case, the ve-locity axis is divided into a series of non-overlapping intervals, preferably the same as theabove-discussed non-overlapping intervals for mapping driving data sets into basic historicgroups. Then, for each such interval, a corresponding allowed engine rotation speed intervalis defined. ln figure 6b, these engine rotation speed intervals are of equal length, but theymay also, for instance, be broader for vehicle intervals for which there are fewer observed previous-trip driving data sets in the collection in question.

Then, a class conformance measure could be constructed and used as follows: 1) For a respective vehicle velocity point P1, [...], P11 in each velocity interval, calculatean average engine rotation speed for the particular vehicle that is to be classified. Thisvalue can be calculated using interpolation in addition to averaging, in case data is notavailable for the particular vehicle velocity value in question. ln this case, the enginerotation speed intervals can be viewed as the class-defining parameters. 2) For each point P1, [...], P11, calculate whether or not the point is within the respectiveengine rotation speed interval. 3) ln case each point Pa, [...], P11, or at least a certain predetermined proportion of thepoints, is or are within the respective engine rotation speed interval, the conformancemeasure turns out in the positive, and the vehicle in question is allotted to the class in question.

From figure 6A, it is clear that the particular exemplifying vehicle represented by points P1,[...], P11 is allotted to the class represented by curve C and the illustrated set of engine rotation speed intervals. lt is realized that many different ways of performing such a conformance measurement be-tween a particular vehicle and a particular collection are thinkable and possible. For in-stance, when there are many classes, one particular vehicle could be found to conform toseveral such classes. ln that case, the conformance measure can further comprise a close- ness measure, based upon which the single class to which the vehicle is closest is the one 37 to which the vehicle is mapped. This closeness measure may, for instance, comprise a meas-ure of engine rotation speed distance to the centre of each engine rotation speed interval for each point P1, [...], P11, or another suitable measure.

According to one preferred embodiment, the class-defining parameters are mutable, and inparticular dynamically updated as new driving data set data becomes available. Hence, it ispreferred that the class-defining parameters for a certain class, preferably for all classes orat least substantially all classes, are automatically and dynamically updated in response tothe collecting of previous-trip data sets, so that the said characteristic engine rotation speedfor a particular vehicle velocity in question is updated in response to the observation andcollecting of a set of instantaneous vehicle speed and instantaneous engine velocity datavalues for a particular vehicle which has been classified into the certain class. This may takeplace by identifying a corresponding basic historic group in the collection, corresponding tothe class in question, to which the previous-trip driving data set in question is mapped; up-dating that basic historic group; and then using the updated basic historic group togetherwith other basic historic groups involving similar vehicle velocity data to update the saidclass-defining parameters. This way, the class definitions will automatically become more accurate as more data becomes available to the system.

When setting up a new system according to the present invention, it may be so that nodriving data set information is available. ln that case, a standard set of initial vehicle classes,as defined by corresponding class-defining parameters can be assumed as a starting point,after which the class definitions may evolve over time as new data becomes available. Al-ternatively, the system uses a basic set of driving data sets, and the initial classes may becalculated based upon the said basic set of driving data sets, and then the class definitions may evolve from there during the use ofthe system. ln either case, and also in other cases, from time to time a vehicle will be observed, thedriving data sets of which are quite far from the closest class definition (as measured by saidclass conformity measure). ln this case, it is preferred that the system may recognize this vehicle as belonging to a new vehicle class and as a reaction create such a new class based 38 upon the driving data sets collected for that vehicle. Thus, in case a particular vehicle isfound to be further away from each of said classes than a predetermined threshold dis-tance, as measured by the said basic class conformity measure, an additional class is cre-ated, together with corresponding class-defining parameters and a corresponding collec-tion. The class-defining pa ra meters ofthe newly created class are then preferably calculatedbased upon the previous-trip driving data sets observed for the vehicle in question. lt isfurther preferred that one single such vehicle, performing one single trip during which it isobserved to be far away from the closest existing class, does not trigger the creation of anew vehicle class, but that at least a certain minimum number of vehicles and/or a certainminimum number of trips is required in order to actually launch the new vehicle class. Thesystem may also comprise limitations for class creation based upon trusted vehicles (seeabove). ln the latter case, a minimum number of trusted vehicles may be required to be observed to belong to a new class before such new class is actually created.

As the number of classes grows, it is preferred that the basic class conformity measure isadjusted correspondingly, so that each existing class more narrowly defines the respectivevehicle, for instance by using more and more narrow velocity and/or engine rotation speedintervals of the type described above, so that the threshold, in terms of distance from theclosest existing class, used as a requirement to launch a new vehicle class, becomes lowerand lower. For instance, interval lengths may be calculated based upon the total number ofclasses and/or the total number of observed previous-trip driving data sets in the database 151.

This way, a more and more granular and fine-tuned set of vehicle class definitions will becreated over time, as more data becomes available to the system, in a way which is fullyautomatic and produces vehicle classes that actually correspond to the main types of vehi-cles that use the system. lt is noted that no a priori knowledge about such vehicles is nec- essary to achieve these results.

Figure 7 illustrates the above described methodology. ln a first step, an initial set of classes is defined. Then, previous-trip driving data sets are collected, for a particular vehicle, but 39 over time for many different trips performed by many different vehicles. For each such ob-served vehicle, the vehicle in question is mapped to the closest existing class, as describedabove and based upon said previous-trip driving data sets and the said class-defining pa-rameters for the respective class. ln case the vehicle was successfully mapped to a class, themapped class is updated, by the previous-trip driving data sets updating the basic historicgroups of the corresponding collection, and the method loops back to collecting previous-trip driving data sets. On the other hand, in case the vehicle was found to be too far fromthe closest class, it is investigated whether or not enough data indicating the motivation tocreate a new class has been collected, as described above. ln case this is so, a new class iscreated, the class-defining parameters of which are based upon the collected driving datasets of the vehicle in question, possibly in combination with previous-trip driving data setsobserved and collected for additional vehicles that are also used for the creation of the newclass. Thereafter, the method loops back again to the collecting of previous-trip driving data sets, using the updated set of class definitions.

Each observed vehicle is preferably mapped to a particular single class before a perfor-mance parameter is calculated for that particular vehicle, and the calculation of the perfor-mance parameter is preferably based only on driving data sets ofthe corresponding collec-tion. However, a reclassification of each vehicle may be achieved less frequently than eachobserved trip, preferably less frequently than every ten trips. However, the updating of theclass-defining parameters of the class to which a particular vehicle is allotted is preferablyperformed at least in connection to the finalizing of each trip performed by the vehicle in question.

According to one preferred embodiment, the class-defining parameters do not compriseinformation regarding vehicle gear used. The surprising finding of the inventor is that gearnumber information does not significantly improve the results, in terms of classification ac-curacy for the purpose of producing relevant driving performance parameter values. lt iseven so that, since the gear usage affects the engine rotation speed for a particular vehiclevelocity, it is difficult to predict suitable class-defining parameter values for a particular ve- hicle, even in the case in which all technical data about the vehicle is known. Hence, in case a particular known vehicle type, such as a newly launched car model of a particular brand,it is preferred that a new vehicle class is created, if needed, automatically by simply con-necting one or several cars of the newly released model, preferably marked as ”trusted”,and then allow the system to automatically discover and define a new set of class-definingparameters for the car model in question, based upon the driving data sets observed for these vehicles.

According to one preferred embodiment, in addition to the above defined collections ofprevious-trip driving data sets and/or basic historic groups, there is a main collection de-fined, comprising respective previous-trip driving data sets, and in particular basic historicgroups corresponding to, such as having identical respective definition as, all respectivebasic historic groups comprised in all of the above described collections. ln this case, a cor-responding basic historic group in the said main collection is always updated with respectto its group performance parameter when a group performance parameter ofa correspond-ing basic historic group in another collection is updated. Hence, data of the basic historicgroups ofthe main collection reflect the data ofall basic historic groups in other collections.According to one preferred embodiment, the data of the basic historic groups in the maincollection can be used instead of basic historic groups for a particular other collection undercertain conditions. For instance, a suitable class may not exist to which the current vehiclecan be mapped, or the current collection may not have enough updated group performance data in order to produce a reliable result.

Fifth aspect According to one aspect ofthe invention, the above-discussed trip performance parametervalue is calculated as a first trip performance parameter value in a way which is similar tothe trip performance parameter value calculation methodology described above in connec-tion to figure 4. ln fact, everything described in connection to figure 4 is relevant also to this aspect ofthe invention, as applicable. 41 This present aspect is further illustrated in figure 9, wherein it is shown that, in a first step the previous-trip and current-trip driving data sets are collected, as described above.

Then, for each of the collected current-trip data sets, at least one corresponding collectedprevious-trip data set is selected based upon the above described basic driving data setsimilarity measure, which is arranged to measure similarity between driving data sets,and/or the above described basic group conformity measure, which is arranged to measureconformity for a current-trip data set to a basic historic group of previous-trip data sets.Hence, according to one embodiment, current-trip driving data sets are selected by map-ping to individual previous-trip driving data sets, and these selected previous-trip drivingdata sets are used for the subsequent calculations. However, it is preferred that the abovedescribed mechanism using basic historic groups, preferably also using the above described classes and collections, is employed.

Thereafter, a relative instantaneous vehicle energy consumption value is calculated for saidselected corresponding previous-trip data set or sets, which relative energy consumption isrelative to a total energy consumption for a respective trip during which the previous-tripdata set in question was observed. lt is understood that, in case the said basic historicgroups are used, the selected previous-trip driving data sets are the ones comprised in thebasic historic groups to which the current-trip driving data sets were mapped. ln particular,it is preferred that the said relative instantaneous vehicle energy consumption value is cal-culated for the respective previous-trip driving data sets in a basic historic group to whichthe current-trip data set in question is mapped, in relation to a total energy consumptionfor the complete trip during which the previous-trip driving data set in question was ob-served, and further preferably based upon an average value of said relative instantaneous energy consumption values for mapped respective basic historic groups. ln a last step, the said first current-trip driving performance parameter value is calculatedbased upon an average value of said calculated relative instantaneous energy consump- tions. 42 Hence, the respective relative energy consumption for each individual previous-trip drivingdata set is a measure of the relative ”goodness”, in terms of low energy consumption, thatthe previous-trip driving data set in question was associated with during the previous tripin question. Similarly, the respective group performance parameter value for each historicgroup, is a measure ofthe ”goodness” that previous-trip driving data sets that have a similarfootprint in terms of basic parameter set values on average are associated with. Then, thefirst performance parameter is a measure of the average such ”goodness” associated withprevious-trip driving data sets, or basic historic groups, which are similar to the collected current-trip driving data sets.

Hence, by breaking a current trip apart into a large multitude of small current-trip drivingdata set observation fragments, associating them with previously observed such fragmentsand calculating the first trip performance parameter value in the way illustrated in figure 9,the above advantages in terms of automatically and accurately assessing comparable driv-ing performance with little a priori knowledge and under shifting conditions are achieved,and the resulting first trip performance parameter value constitutes an easily accessible,numerical value that is directly useful as a trip performance measure. Hence, the trip per-formance parameter value can be displayed to the driver during or after the current trip, asdescribed above, but it can also readily be used for making direct comparisons betweendifferent trips and drivers, and even between different vehicles, since the first trip perfor-mance parameter is generally independent of driver and driving conditions. ln particular, incase the mechanism using classes and collections described above is used, the first perfor-mance parameter value will also be generally independent upon vehicle type, so that a trip using a bus is readily comparable to a trip using a small car.

The portable electronic device 130, or alternatively the current vehicle 100 itself, is prefer-ably arranged with a piece of software arranged to present to the driver of the current ve-hicle a graphical user interface, in turn arranged to present information comprising a repre-sentation ofthe calculated first trip performance parameter value for the current trip, andpossibly also for previously conducted current trips for the same driver, and possibly also for other previous trips. 43 The software of said portable electronic device 130 may be a piece of software executableby or from the portable electronic device 130, such as a locally installed and executed ap-plication, a remotely executed application, such as a web page application accessed from the portable electronic device 130, or any other suitable type of software. ln the preferred case in which the first trip performance parameter is calculated repeatedly,by central sever 150 or local server 160, during the current trip, based upon the so far col-lected current-trip driving data sets, it is preferred that a representation of an updated firstparameter value is presented to the driver on said graphical user interface during the cur- rent trip. lt is furthermore preferred that the calculated first trip performance parameter values cal-culated for previous trips are stored in the database 151 and are made available via a suit-able application programming interface (API) provided by the central server 150. This way,the manager of a fleet of transport vehicles, or similar, can follow the progression of thefleet in terms of driving performance over time, and perform analyses based upon first trip performance parameter value data for the fleet. ln a particularly preferred embodiment, the first performance parameter is used to calcu-late a benchmark value. Once a plurality of first performance parameter values have beencalculated, preferably for a plurality of different trips with a plurality of different vehiclesand by a plurality of different drivers, the system determines a threshold first trip perfor-mance parameter value such that only a minor percentage, such as 10%, of all calculatedtrip performance parameter values, are better than the said threshold value. Then, eachnewly calculated first trip performance parameter value can be compared to the benchmark value to see how far from the 10% top previous trip performances that the current trip was. lt is furthermore preferred that the benchmark value is updated based upon first trip per-formance parameter values calculated with respect to the current trip, at least as long as the benchmark value has not converged so that it substantially does not change with the 44 updating of newly calculated trip performance parameter values. lt is preferred that the benchmark value is made accessible throughout the system as a global variable.

Sixth aspect ln one aspect of the invention, the value of a second trip performance parameter is calcu-lated, based upon several previously calculated trip performance parameter values, prefer-ably but not necessarily several previously calculated trip performance parameter values of the above explained type, namely the first trip performance parameter value. ln general, the second trip performance parameter value is calculated based upon the samedata as the first trip performance parameter value, in terms of previous-trip and current-trip driving data sets, basic historic groups, vehicle classes, collections, etc., as described indetail above. However, the idea behind the second trip performance parameter value ismore generally applicable than that of the first trip performance parameter value. ln par-ticular, it is not strictly necessary, albeit preferred, that the basic parameter set, for thepurposes of calculating the second trip performance parameter value, comprises instanta-neous vehicle energy consumption. lnstead, some other parameter or combination of pa-rameters comprised in the basic parameter set can be used to in some respect measure therelative quality of each previous-trip driving data set, such as parameters measuring tyrewear (for instance a suitable parameter combination of observed break usage and turningmagnitude in relation to vehicle velocity). Such non energy consumption-based parametersmay be used in case the measurement aim is different from the one described in detailherein below, but in case the basic mechanism of the second trip performance parameter calculation, and its general advantages are still desired.

For reasons of simplicity, in the following the calculation of the second trip performanceparameter will be described as if the basic parameter set comprises instantaneous energyconsumption. ln general, everything which is said in relation to the first trip performanceparameter herein is equally useful for the purposes of calculating and using the second trip performance parameter value.

Figure 10, which is similar to figure 9, illustrates the basic methodology for calculating thesecond trip performance parameter value according to the present aspect of the invention.ln a first step, current-trip and previous-trip driving data sets are collected, as describedabove, and for individual current-trip driving data sets corresponding previous-trip drivingdata sets are selected. Thus far in the method, this aspect is in many regards the same as for the aspect described above in connection to figure 9.

However, for each selected previous-trip driving data set, a quality measure is then calcu-lated. This quality measure may be the above described relative energy consumption-based performance measure, but may also be something else.

Thereafter, a respective first trip performance parameter value is calculated for each previ-ous-trip data set, which first trip performance parameter may be the same as the above-described first trip performance parameter. However, it may also be another suitable typeoftrip performance parameter, the value ofwhich is calculated based upon the said qualitymeasure. Preferably, the first trip performance parameter is a relative trip performance pa-rameter arranged to measure the relative trip performance ofthe previous-trip driving dataset in question in relation to the trip during which the previous-trip driving data set wasobserved. ln the exemplifying case in which the quality measure is a combination of instan-taneous break and turn data, the first trip performance parameter for each previous-tripdriving data set may be a relative value forthis instantaneous break/turn parameter as com-pared to an average value of said parameter for the complete trip during which the previ- ous-trip driving data set in question was observed. ln a final step, the second trip performance parameter value is calculated based upon therespective values of the said first trip performance parameter for each of the selected pre- vious-trip data sets. 46 Figure 11 illustrates a method according to the present aspect of the present invention, inparticular in which the above described basic historic groups are used for the calculation of the second trip performance parameter.

Hence, in a first step, for many previous trips, such as for at least 100 previous trips, prefer-ably at least 1000 previous trips, respective previous-trip driving data sets are collected. Foreach such previous-trip driving data set, a relative quality measure is calculated, preferablyrelative to a total quality for the complete previous trip during which the previous-trip driv-ing data set in question was collected, in particular preferably the above described instan- taneous relative energy consumption. ln a second step, each previous-trip driving data set is mapped to a basic historic group. ln a third step, for each such mapped basic historic group, a respective group performanceparameter value, preferably the above described energy consumption-based group perfor- mance parameter, is updated using the calculated relative quality measure. ln a fourth step, a first trip performance parameter value, preferably the above describedenergy consumption-based one, is calculated for each of said many previous trips, based upon the updated group performance parameters for the mapped respective basic groups. lt is noted that, in this fourth step, each previous trip can be regarded as a current trip, andthat the first trip performance parameter value then corresponds to the above described trip performance parameter calculated for a current trip. ln a fifth step, again for all of said many previous trips, a respective general group perfor-mance parameter is updated for each of the respective mapped basic historic groups cor-responding to the previous-trip driving data sets observed during the previous trip in ques-tion, which update is based upon the updated respective first trip performance parametervalue calculated in the fourth step. Preferably, the general group performance parameter is an average value, such as a geometric average, of the respective first trip performance 47 parameter values previously calculated for all previous-trip driving data sets mapped to the basic historic group in question. ln a sixth step, current-trip driving data sets are collected for a current trip. ln a seventh step, each such current-trip driving data set is mapped to a respective basic historic group, in the way described above.

Then, in an eighth step, the second trip performance parameter value is calculated basedupon the respective general group performance parameter values calculated for all basichistoric groups to which current-trip driving data sets are mapped. Preferably, the secondtrip performance parameter value is calculated as an average value, such as a geometricaverage value, of the said general group performance parameter values. The said averagevalue can also be a weighted average value, such as an average value in which more fre-quently updated basic historic groups are given more weight than less frequently updatedbasic historic groups. ln case not all current-trip driving data sets correspond to a respectiveexisting basic historic group, the averaging function may ignore those current-trip driving data sets for the purpose of calculating the second trip performance parameter value.

Using the present system and method for calculating, for a current trip, said second tripperformance parameter in the above described way achieves the surprising effect that thevalue of the second trip performance parameter constitutes a very accurate measure of therisk level assumed by the current driver. ln other words, the second trip performance pa-rameter measures the risk behaviour of the driver. Apart from this aspect, all the ad-vantages described above, in relation to the calculation ofthe first trip performance param- eter, also apply to the second trip performance parameter. lt is preferred that the second trip performance parameter is made available to the currentdriver after or during the current trip in a way which completely corresponds to the case forthe first trip performance pa ra meter, as described above. lt is also preferred that the second trip performance parameter values for individual drivers and/or collectives of drivers are 48 used for evaluation and risk assessment purposes. For instance, an insurance company mayuse the value of the second trip performance parameter as an input in the calculation of carinsurance premiums. Furthermore, such second trip performance parameter value may beused to identify risk-assuming individuals or groups of drivers for the purposes of improvingthe operations of a transport company. There are numerous other ways in which such a measure of risk can be used.

According to one preferred embodiment corresponding to figure 11, the collected previous-trip data sets are classified into one of a plurality of different predetermined basic historicgroups based upon said basic similarity measure, and each current-trip data set is mappedto at the most one of said basic historic groups based upon said basic group conformitymeasure. Then, each of the previous-trip driving data sets are further mapped to at themost one of said basic historic groups, at the time comprising previous-trip driving data setsobserved before the previous-trip driving data set in question was observed, based upon said basic group conformity measure. lt is understood that the above-described preferred case in which the first (and conse-quently also the second) trip performance parameter is calculated based upon a measureofthe instantaneous energy consumption ofthe previous vehicles, the qualified driving dataparameters of the previous-trip driving data sets comprise instantaneous energy consump-tion. Then, the method comprises a step in which a relative instantaneous vehicle energyconsumption value is calculated for said previous-trip driving data sets, which relative en-ergy consumption is relative to a total energy consumption for a respective trip duringwhich the previous-trip driving data set in question was observed. Furthermore, in this casethe first trip performance parameter value is calculated based upon such calculated relative energy consumption values. ln particular, in this case it is preferred that, for the respective basic historic group to whicheach current-trip driving data set is mapped, and for each further basic historic group towhich each previous-trip driving data set comprised in the basic historic group in question is in turn mapped, a respective relative instantaneous vehicle energy consumption value is 49 calculated, which relative energy consumption is relative to a total energy consumption fora respective trip during which the previous-trip driving data set in question was observed.Furthermore, in this case each of said first trip performance parameters is calculated based upon an average value of said relative instantaneous energy consumption values. lt is noted that, in case the methodology with vehicle classes and collections describedabove is used, all calculations leading up to the second trip performance parameter value are limited to the current collection.

Seventh aspect ln one aspect of the invention, illustrated in figure 14, current-trip and previous-trip drivingdata sets are collected in a first step. Specifically, updated current-trip driving data sets arerepeatedly read from the vehicle, wherein new such current-trip driving data sets are readfrom the vehicle at consecutive observation time points separated by at the most a prede-termined observation time period. This is similar to the above described aspects of the pre-sent invention. However, the current-trip driving data sets in the present aspect each com- prises data from at least a predetermined set of extended driving data parameters.

Furthermore, in the present aspect, the previous-trip driving data sets also comprise thesaid extended driving data set parameters, and in addition thereto the previous-trip drivingdata sets each comprises parameter values for a predetermined set of a qualified parame-ters. The qualified parameter set, in turn, comprises the parameters of the above defined basic parameter set as well as, or comprising, instantaneous vehicle energy consumption.

Hence, the qualified parameter set at least comprises the basic parameter set. ln case thebasic parameter set does not comprise instantaneous vehicle energy consumption, thequalified parameter set adds this parameter as compared to the basic parameter set. Theextended parameter set, in turn, comprises parameters that may or may not have an over- lap with the qualified parameter set, as explained below. To sum up, in this aspect each previous-trip driving data sets comprises value for the basic parameter set as well as addi- tional information.

The possible order ofthe various steps in this aspect is illustrated by arrows in figure 14.

Hence, in a second step, the co||ected previous-trip driving data sets are a||otted to basichistoric groups, as described above using the said basic driving data set similarity measure operating on the basic parameter set values of the previous-trip driving data sets. ln a third step, the value of a respective group performance parameter is then calculatedfor each basic historic group, in a way that may be as described above. Specifically, thegroup performance parameter may, but needs not, be a relative energy consumption-basedperformance parameter as described above. lt is preferred that the group performance pa-rameter value is calculated based upon only the qualified parameter set for each basic his- toric group. ln a fourth step, the said previous-trip driving data sets are grouped into a set of historicextended groups of previous-trip driving data sets, such that each previous-trip driving dataset is a||otted to one such extended historic group. Like in the case for the above describedbasic historic groups, each such extended historic group is a group of previous-trip drivingdata sets. However, in contrast to the case for the basic historic groups, previous-trip drivingdata sets are a||otted to one of said set of extended historic groups based upon an extendeddriving data set similarity measure, which extended similarity measure is arranged not totake all values for said basic parameter set into consideration that are taken into consider-ation by the basic similarity measure. An extended group conformance measure may alsobe used, which then corresponds to the above described basic group conformance meas- Ufe.

That the extended similarity measure is arranged not to take all values for the basic param-eter set into consideration that are taken into consideration by the basic similarity measure means that at least one parameter of said basic parameter set is not used for calculating the value of the extended similarity measure for the purposes of grouping together previ-ous-trip driving data sets in extended historic groups. However, it is preferred that none ofthe parameters in the basic parameter set is used for such calculation. Nevertheless, a cer-tain parameter which forms part ofthe basic parameter set can be a parameter measuringthe same thing but in a different way, and hence count as not the same parameter for thesepurposes. For instance, even if the basic parameter set comprises vehicle speed, and theprevious-trip driving data sets comprises such a parameter value, measured on the actualvehicle engine or the wheel axis, the extended parameter set may also comprise vehiclespeed, and the previous-trip driving data sets may then also comprise a vehicle speed valueas measured using a GPS component in the portable electronic device 130. lt is noted that,even though these parameter values correspond to the same metric, they are in general notnumerically the same, and are subject to different artefacts and error sources. There arenumerous other examples in which a certain metric can be measured both directly on vehi-cle hardware and in some other way, for instance using sensors of the portable electronicdevice 130 such as GPS, accelerometer, gyro, compass, etc. components. Hence, the meas- urement method may be part of the definition of a "parameter". ln a fifth step, for each of said collected current-trip data sets, the current-trip data set inquestion is mapped to at the most one particular one of said extended historic groups,based upon an extended group conformity measure between a driving data set and an ex- tended historic group.

The extended group conformity measure may be similar to the above discussed basic groupconformity measure, in that it may, for instance, use predefined intervals for the extendedparameter set values and allot a certain current-trip driving data set to a certain extendedhistoric group in case all extended parameter values of the current-trip driving data set fall within the corresponding interval ofthe extended historic group in question. ln a sixth step, a current-trip driving performance parameter is calculated based upon the said group performance parameter values calculated for each respective basic historic group corresponding to the previous-trip driving data sets comprised in the extended his- toric group to which a current-trip data set was matched in the fifth step.

Hence, the extended historic groups, that are used in parallel to the basic historic groups toclassify co||ected previous-trip driving data sets, but that use the extended parameter setto perform the classification as opposed to the basic parameter set, constitutes a link be-tween data measured on the vehicle and externally measured data or data read otherwisenot in direct contact with the vehicle as such. By mapping current-trip driving data sets toextended groups, it is possible to use the information represented by previous-trip drivingdata sets in the same or corresponding way as described above, even in case the current-trip driving data sets do not comprise the basic pa rameter set and are therefore not possibleto map to a particular basic historic group based upon the above described basic similarity OI' COHfOFmanCe meaSUFe. lt is even preferred, in the present aspect of the invention, that the current-trip driving datasets do not comprise said basic parameter set, at least to the extent to which sufficient datais lacking such that using the above described basic similarity and/or conformance measuresbecomes impossible. ln some embodiments, it is enough that only one basic parameter is lacking from the extended parameter set for such use to be impossible.

Figure 15 is a more detailed view of an exemplifying embodiment of the present aspect.

Firstly, previous-trip driving data sets are co||ected for many previous trips, and a relativequality (such as an energy consumption-based quality in relation to a corresponding com- plete trip) is calculated for each co||ected previous-trip driving data set, as described above.

Each co||ected previous-trip driving data set is mapped to a particular one basic historicgroup, based for instance upon the said basic conformity or similarity measure, and therespective basic group performance parameter is updated using the calculated relative quality value.

Furthermore, each collected previous-trip driving data set is mapped to a particular oneextended historic group, based upon said extended similarity or conformance measure. Foreach such mapped extended group, a corresponding extended group performance param-eter value is updated using a corresponding basic group performance parameter value taken from the basic group to which the previous-trip driving data set was allotted.

Recalling that the basic group performance parameter may be a, possibly weighted, averagevalue of the relative quality measures calculated for each previous-trip driving data set al-lotted to the basic historic group in question, the extended group performance parametermay, in a corresponding way, be an average value ofthe corresponding basic group perfor-mance parameter values used to calculate the extended group performance parameter. lnparticular, it is preferred that the extended group performance parameter is a weightedaverage value, wherein more frequently used basic historic groups are given larger weight than less frequently used basic historic groups.

As a result, each extended group performance parameter will, over time as the system isused, become a measure of average relative driving quality for the basic historic groups towhich the same previous-trip driving data sets were allotted as were allotted to the ex- tended historic group in question.

When the current-trip driving data sets are then collected, they are each mapped to a re-spective one extended historic group, and the respective extended group performance pa-rameter values ofthe mapped extended historic groups are used to calculate a trip drivingperformance parameter value, which is then used as the trip performance parameter values described above.

According to a present invention, the extended parameter set comprises at least one pa-rameter from a parameter list comprising GPS-based velocity, GPS-based acceleration, alti-tude, accelerometer-based acceleration and compass-based heading. The said list may also comprise corresponding changes over a predetermined time period, in a way which corre- sponds to the above described insta ntaneous vehicle velocity and insta ntaneous vehicle ve-locity change, as well as to instantaneous engine speed and instantaneous engine speed change. lt is furthermore preferred that at least one of said current-trip driving data set values, pref-erably all of the current-trip driving data set values, are either registered by the currentvehicle, which vehicle is connected to a central server via a wireless connection, or, evenmore preferably, registered by a portable device arranged in the current vehicle, which portable device is connected using a wireless connection to the said central server.

This way, a current vehicle which itself has no capability of recording data corresponding tothe basic parameter set can still be used with the system, by recording data correspondingat least to the extended parameter set, and then receiving a trip performance parametervalue which draws upon the total previous-trip driving data set pool collected for previousvehicles that in fact did have basic parameter set data reading capabilities. The only require-ment is that such previous vehicles also recorded extended parameter set data during theprevious trips, so that it was possible to map the previous-trip driving data sets to appropri- ate extended historic groups. ln order to achieve the latter, it is preferred that the software of the above-described pieceof portable electronic device 130, providing graphical user interface is arranged to measurethe complete extended parameter set data during each current trip, and to report the meas-ured extended parameter set data to the central server 150 for processing. Such measure-ment is preferably performed using sensor data locally available to such piece of software,preferably using sensor hardware integrated into the portable electronic device 130, evenif it could also be measured by the vehicle itself, or by device 120. This way, a softwareservice used by the users of the present system can automatically record extended param-eter set data, preferably in addition to basic or qualified parameter set data, in a way which is completely transparent to the user, for use by other current vehicles using the system. ln particular, it is preferred that the current vehicle is not arranged to automatically providedriving data information via an external interface. ln this case, the present aspect is namelyparticularly useful. For instance, the present method can be used in a car without any suchexternal interface, or when a required piece of hardware 120 is lacking or broken, by simplyusing a smartphone of the user or similar. The present invention is even useful for measur-ing driving performance for non-motorized vehicles, such as bicycles, as long as enough rel-evant previous-trip driving data sets with a relevant ”driving quality” measurement have been recorded covering the currently used basic parameter set.

Hence, different systems may be implemented with particular adaptations to suit particularvehicle types, wherein relevant selections are made with respect to basic, qualified and ex-tended parameter sets, and the notion of ”driving quality”. For instance, ”driving quality” for a bicycle with an electrical help motor could be related to the use of battery pack power. lt is furthermore preferred that only trusted vehicles, as described above, are allowed to update the said extended group performance parameter values. This will improve data qual- ity. ln case collections are used, as described above, it is preferred that the basic historic groupsused for the purposes of calculating the extended group performance parameter are taken from the above described main collection.

General ln general, it is preferred to use basic historic groups, as described above, that are used tostore information regarding previous-trip driving data sets mapped to such basic historicgroups. Hence, it is preferred to, as a part of each basic historic group, store and update notonly the group performance parameter, but also the above described trip performance pa-rameters for current trips during which a current-trip driving data set was mapped to thebasic historic group in question, notably the first trip performance parameter. Such param- eter data which relates to the basic historic group as such is preferably updated dynamically as an average value of incoming data. For instance, this may be achieved by the currentlyupdated parameter value being stored in one memory position in the database 151, andthe number of previous updates in an additional memory position in the database 151.Then, as a new update arrives, the |atter number can be increased by one, and the basic historic group parameter can be updated according to the following, as an example: _ PN'N+P PN+1_ N+1 r wherein N is the number of previous updates; P is the stored, average parameter value; and p is the incoming, new parameter value.

Apart from such average parameter values, it is furthermore preferred to store additionalinformation for each basic historic group. One example is when a traffic accident occurs.An accident can be confirmed in various ways, such as by automatic detection based upondriving data sets, such as a rapid decrease in vehicle velocity followed by a standstill, or bymanual registration. Once an accident has been confirmed, the system is preferably ar-ranged to collect a predetermined number of current-trip driving data sets, and for eachbasic historic group to which the collected current-trip driving data sets are mapped, suchas using the basic group conformance measure, update an accident risk parameter valuestored for each such basic historic group. This update can be a simple counter which is in-creased by one each time this occurs for the basic historic group in question, or it may bean average value updated as described above. Hence, such an accident risk parameter,when used in the system over a prolonged time period during which a number of con-firmed accidents occur, will be a measure ofthe probability of each basic historic groupbeing observed during a trip leading to an accident. Hence, a separate trip performanceparameter value can be calculated based upon the accident risk parameter values forbasic historic groups to which current-trip driving data sets are mapped during a currenttrip, or the above described second trip performance parameter value can be calculatedbased at least partly upon such accident risk parameter values, in addition to the abovedescribed calculation, with the result of a trip performance parameter being accomplishedwhich more accurately takes into consideration driving behaviour which is known to be risky.

Furthermore, the above described calculation ofthe second trip performance parameterduring the current trip may be used to, during the current trip, predict a high risk of acci-dents for the current trip, based upon a poor value ofthe currently calculated second tripperformance parameter. ln this case, the system is arranged to provide a warning to the current driver.

This latter is made possible by the general property of a system according to the presentinvention to statistically relate driving effects on a macroscopic time scale to causes interms of the driving behaviour on a microscopic time scale. This is a key insight of the pre- sent inventor. lt is noted, in connection to this calculation ofthe said accident risk parameter value, thata pattern of mapped basic historic groups is preferably not identified and stored as such inconnection to an accident and for the purposes of identifying particularly accident-pronepatterns; instead, individual information is stored for each individual basic historic group,and a trip performance parameter value is then calculated based upon the individual basichistoric groups and the said data. Since the number of basic historic groups is large, pref-erably at least 100.000 basic historic groups, more preferably at least 1.000.000, the re-sulting trip parameter value will still in general be an accurate measure ofthe metric beingmeasured, as applicable. This is true in general for all the above-described aspects of the present invention.

The longer the system is used, the more data, in terms of updated basic historic groups,each on average having been updated with many previous-trip driving data sets, the data-base 151 will contain. As a result, resulting calculated trip performance parameters willbecome more and more accurate, even for current vehicles that have previously not beenconnected to the system. Hence, the system is, in this regard, a ”learning” system in the sense that it is automatically improved during use.

Regarding the above-described trusted vehicles, according to one preferred embodiment there is a parameter in the system which is settable so that certain selected trusted vehicles are associated with an increased weight when updating system data. Then, such trustedvehicles can be used to quickly adapt the system, in terms of vehicle class definitions (class-defining parameter values) and basic historic group data (in particular group performanceparameter values), when for instance new car models are launched. ln one preferred exam-ple, such selected trusted vehicles are associated with an increased weight of at least 5,preferably at least 10, times a default weight of a trusted vehicle. Hence, when previous-trip driving data sets are collected for such a trusted vehicle, this data counts as at least 5,prefera bly at least 10, such previous-trip driving data sets collected at the same time, andall parameter value updates are performed using this weight. As described above, this maylead to a new vehicle class automatically being created for such a newly released car model,but it may also lead to an existing vehicle class definition being adapted to the new carmodel, depending on how different the new car model is from already observed previousvehicles. ln one preferred embodiment, such higher-weight updates only apply to class-de- fining parameters, and not to, for instance, group performance parameter values.

According to one preferred embodiment, each basic historic group is associated with a dataquality parameter value, indicating the data quality of the group performance parameterfor the basic historic group in question. According to one embodiment, such data qualityparameter can indicate whether or not a previous-trip driving data set has been mapped tothe basic historic group in question. ln case this is not so for a particular basic historic group,the group performance parameter may be excluded from the calculation of the above de-scribed trip performance parameters described above. This may, for instance, be accom-plished by the basic historic group in question being ignored for the purposes of such calcu-lation. ln case fewer than a predetermined percentage of the observed current-trip drivingdata sets can be mapped to a respective basic historic group the data quality parameter ofwhich indicates a less than full quality, this may be indicated by the system, for instance bydisplaying a warning to the user of the current vehicle indicating that the calculated tripperformance parameter is of potentially poor quality. The said predetermined percentageis preferably between 50% and 90%. Alternatively, the relative percentage of full-qualitybasic historic groups can be used to calculate a confidence interval with respect to a calcu- lated trip performance parameter, and then displayed to the current driver.

Once a previous-trip driving data of a trusted vehicle set is mapped to the basic historicgroup in question, the data quality parameter may be updated, so as to reflect a higherstate of data quality. According to one preferred embodiment, the group performance pa-rameter is updated even before this happens, based upon previous-trip driving data sets ofnon-trusted vehicles that are mapped to the basic historic group in question. Then, oncethe data quality parameter is set to indicate a higher data quality, the hence updated groupperformance parameter will become available for use when calculating trip performanceparameters. This approach has turned out to provide accurate trip performance parameter values while still keeping the system simple yet dynamically adaptive.

Furthermore, as described above, the basic historic groups of different collections may beupdated as a result of collected previous-trip driving data sets for vehicles of different vehi-cle classes. This, in turn, will in general lead to different collections comprising differentlyfrequently updated group performance parameters, and data of different collections hencehaving different data quality. ln this case, it is preferred for the system to comprise func-tionality for periodically investigate whether basic historic groups have data quality param-eters that are set to indicate higher data quality, and for which corresponding basic historicgroups of other collections do not have data quality parameters that are set to indicatehigher data quality. |fthis is found to be the case, the respective group performance param-eter values of basic historic groups with lower data quality may be updated using respectivegroup performance parameter values of corresponding basic historic groups, in other col-lections, with higher data quality. ”Corresponding” basic historic groups, in this context,preferably means basic historic groups with identical definition. The update is in this casepreferably performed as a weighted average calculation ofthe lower-quality group perfor-mance parameter, wherein the weight of the higher-quality group performance parameteris lower than the weight of the lower-quality group performance parameter. Preferably,such updates between different collections only takes place between collections the corre-sponding classes of which are more similar than a predetermined value, which similarity is measured and calculated using a certain vehicle class similarity measure. This class similarity measure is arranged to measure similarity between two vehicle classes based upon the re- spective class-defining parameters ofthe classes in question.

A similar method can be used when a new class is defined. ln this case, a set of basic historicgroups can be copied from the collection corresponding to another class which is sufficiently”close” the newly created class to the collection corresponding to the new class, which setof basic historic groups has full data quality as indicated by said data quality parameters. lnthis case, it is preferred that the respective group performance parameters of such copiesin basic historic groups are given less weight than normal during data updates in the collec-tion corresponding to the newly created class, so that the convergence of the collection isquicker as driving data sets are being collected for vehicles of the newly created class. Fur-thermore, it is preferred that a characteristic vehicle velocity function is copied from such a”close” class, and used for the newly created class. Thereafter, the copied characteristic function will be updated by collected driving data sets for the newly created class vehicles.

As described above, it is preferred not to store a pattern of basic historic groups to whichdriving data sets have been mapped during a trip resulting in an accident. However, there are cases in which a pattern of mapped basic historic groups is identified, and even stored.

One such example is for driver identification. lt has turned out that the basic historic groupsto which current-trip driving data sets are mapped for each particular driver follows a sta-tistical pattern which may be different enough between drivers so as to be used for driveridentification. Hence, according to one preferred embodiment, the driver of each previousvehicle is identified, and a respective statistical pattern of mapped basic historic groups,corresponding to collected previous-trip driving data sets, is identified and stored for eachdriver, for several previous trips made by each such driver. Then, the current driver can beidentified by a statistical comparison between the stored statistical patterns for the usersand the pattern of mapped basic historic groups during the current trip. Such comparisoncan be performed in any manner which is conventional as such, and would typically resultin one of said stored statistical patterns that represents the best match to the pattern pro- duced during the current trip. 61 A ”pattern”, as used herein, may comprise information both regarding the identity ofmapped basic historic groups; data contents of mapped basic historic groups; and/or map-ping frequency of basic historic groups; or any combination of such parameters. lt is pre-ferred, when determining such a pattern, that data from several previous trips, such as atleast 20 previous trips, of the same driver are used for such determination; and also thatthe previous-trip driving data sets for such trips are filtered so as to remove outlier data points. ln particular, in one preferred embodiment, such driver identification may be used to auto-matically stop a vehicle, or set off an alarm, if a driver which has not been previously au-thorized to drive the current vehicle drives the current vehicle. To this end, the system maycomprise a piece of hardware in the current vehicle arranged to stop the vehicle in a suitable way, such as after providing repeated warnings to the driver. ln another preferred embodiment, driver identification data is stored in the central data-base 151 for all or some previous trips, and may be used to retroactively map particulardrivers to particular previous trips, for instance for the purpose of automatically updatingdrivingjournals, to produce driving statistics or to investigate who drove a particular vehicleduring a particular previous trip for insurance purposes. The stored data may also, for in-stance, be used to verify that a particular person is actually the driver in a motor competi- tion or similar. ln one preferred embodiment, a pattern for a particular driver is used irrespectively of which vehicle and which vehicle class is used during the current trip. ln an application similarto the above described pattern determination, basic historic groupsmapped by previous-trip driving data sets observed for the driver in question are analysed,and it is identified in what basic parameter set intervals, such as in what velocity intervals,the group performance parameter values of the said mapped basic historic groups corre- sponding to those intervals are lowest. Then, this interval information is presented to the 62 user, and used to direct the attention ofthe user to certain fields of improvement regarding the user's driving skills. ln one preferred embodiment, the current vehicle does not have the capability to produceuser-readable fuel-consumption data during the current trip. ln this case, the system is ar-ranged to calculate the fuel consumption for the current trip based upon relative energyconsumption-based group performance parameter values for basic historic groups to whichcurrent-trip driving data sets are mapped. This calculation is straight-forward but dependson the detailed implementation of said performance parameter. The present inventor hasdiscovered that such calculated fuel consumption may be surprisingly accurate, even in thecase in which the basic parameter set does not comprise fuel consumption and when there is no fuel consumption value available for readout from the current vehicle. ln one preferred embodiment, the current driver is an automated driver, such as a software- and/or hardware implemented robot or autopilot.

Further applications for the present invention is to assessing how difficult it is to drive acertain stretch of road, in relative terms as compared to other stretches of road and withrespect to energy consumption and driving riskiness, by performing a number of trips alongthe road in question and noting an average first and/or second trip performance parametervalue for such trips in relation to average corresponding trip performance parameter values for other stretches of road.

Example Figures 16 and 17 illustrate an example of an embodiment of the present invention, for a more detailed understanding ofthe same. ln figure 16, a series of observed and collected previous-trip driving data sets are shown(figure 16, top) along a time axis. The driving data sets are observed at consecutive time points, one second apart, starting at time = 102 seconds. Each previous-trip driving data set 63 comprises the following data values, which are made available by the previous vehicle dur-ing driving and for readout as described above, and communicated to the central server 150 (and/or to the local server 160, as the situation may be): 0 A predetermined qualified parameter set comprisingo A predetermined basic parameter set, in turn comprising I lnstantaneous vehicle velocity, ”Vel” (km/h) I lnstantaneous engine rotation speed, ”RPM” (RPM) I lnstantaneous vehicle velocity change as measured from the obser-vation point and forwards 5 seconds, ”AVel” (km/h) I lnstantaneous engine rotation speed change as measured from theobservation point and forwards 1 second, ”ARPM” (RPM) o lnstantaneous fuel consumption, ”FC” (liters per 10 km) As seen in figure 16, the vehicle velocity measured by the previous vehicle increases, fromtime point 102 to time point 107, from 82 to 85. At the same time, the engine rotation speedincreases from 1550 to 1750. These shifts are also reflected in the change parameter values.ln is realized that the actual processing of the previous-trip driving data set observed at time102 will not actually be performed by the central 150 or local 160 server until time 106, when the velocity change is known.

Simultaneously as the above data is collected from the previous vehicle itself, previous-tripextended driving data sets (see figure 16, bottom) are also collected by a smartphone 130held in the previous vehicle by the driver (in this exemplifying embodiment). The observa-tion time points are identical (time points 102-107, with 1 second apart), but the extendeddriving data observed, collected and communicated to the server in question comprises the following data values: 0 A predetermined extended parameter set comprisingo lnstantaneous GPS-based velocity, ”GPS-Vel” (km/h) o lnstantaneous altitude, ”Alt” (meters above sea level) 64 o lnstantaneous GPS-based velocity change as measured from the observationpoint and forwards 5 seconds, ”AVel” (km/h)o lnstantaneous altitude change as measured from the observation point and forwards 1 second, ”AAlt” (meters) lt is noted that the extended pa ra meter set does not comprise insta ntaneous fuel consump- tion.

Hence, in a step A, the previous-trip driving data set at time point 102 is co||ected and com-municated to the central server 150. ln a step B, the driving data set in question, or moreprecisely, the velocity- and engine rotation speed data comprised in the data set, is used toupdate a current characteristic engine rotation speed to velocity curve for the present classof vehicles. ln a step C, the class conformance measure is used to map said characteristiccurve to one particular of a set of available and dynamically updated vehicle classes. ln stepD, the identified class is used to find a corresponding collection, among several such collec-tions (displayed as circles in figure 16), each corresponding to a particular one of said vehicleclasses. ln the present example, the vehicle class may cover, for instance, middle-sized sta-tion wagon cars. lt is noted that the present system has grouped these vehicles together insuch a class completely automatically, without any presupposed knowledge about how to group vehicles or vehicle properties.

Then, in a step E, the previous-trip driving data set in question is mapped, using the groupconformance measure, to a corresponding basic historic group, among many such availablegroups in said collection. ln the present example, the velocity and velocity change valuesare measured in 1 km/h intervals; and the engine rotation speed and engine rotation speedchange are measured in 50 RPM intervals, which is the same as used for basic historic groupdefinitions, which is hence also based upon the same interval sizes. Therefore, the mappedbasic historic group contains the same basic parameter set data as the previous-trip drivingdata set. This provides for very rapid lookup functionality in the system, in particular when using said classes and collections. ln a later performed step F, the previous-trip driving data set observed at time 103 is pro- cessed in a similar way, and is mapped to another basic historic group. it is preferred that the vehicle is not allowed to change vehicle class during a trip. As a result, step B may be performed for all previous-trip driving data sets at a single, later time.

After step E, in a step G, the mapped basic historic group in question is updated with respectto its group performance parameter (GPP) value. This step is in fact taken after the previoustrip is finished, or alternatively, if performed during the previous trip, under the assumptionthat the previous trip up to the time 102 is the total previous trip. For the total previous tripthen, the total average fuel consumption is read from the previous vehicle, and the instan-taneous fuel consumption, namely 7.5 liters per 10 km, for the previous-trip driving data setin question is divided by the said total average fuel consumption. The result is a percentagevalue, indicating the relative fuel consumption at the small time window of 1 second at timepoint 102, as compared to the whole trip. ln this case, the average fuel consumption for theprevious-trip driving data set in question was lower than previously noted (on average) forthat particular basic historic group in that particular collection, why the GPP of the basicgroup is decreased from 105.31 to 105.27, meaning that the average relative fuel consump-tion for previous-trip driving data sets previously mapped to that basic historic group is now 105.27%.

This is performed for all previous-trip driving data sets of the previous trip in question, or atleast intermittently and under the assumption that the previous trip up to a particular pre-vious-trip driving data set constitutes the total previous trip. Then, in a step H, the respec-tive GPP values for all mapped basic historic groups for the total previous trip are summed,and an average GPP value is calculated for the trip. This value, which is the first trip perfor-mance parameter, is communicated to the device 130 and presented to the previous driver,preferably in relation to a benchmark value for first trip performance parameters as deter-mined based upon corresponding calculations for previously performed previous trips (see figure 17). 66 ln a step I, the general group performance parameter (GGPP) of each mapped basic historicgroup is also updated, using the said calculated first trip parameter value. ln this case, thefirst trip performance parameter value turned out to be 104.85, which is higher than theGGPP (98.11) of the basic historic group in question. Hence, its GGPP value is averaged upto 98.13. The corresponding is done for all mapped basic historic groups. Then, the secondtrip performance value is calculated by averaging all GGPP values for all mapped basic his-toric groups, and the second trip performance parameter is also communicated to the de- vice 130 for display to the previous driver (figure 17).

Also, in a step J, for each mapped basic historic group, a main collection basic historic groupis also identified, using the group conformance measure, and its GPP and GGPP measuresare updated in a way which corresponds to the mapped basic historic groups of the collec-tion corresponding to the vehicle class to which the previous vehicle belongs. lt is notedthat the mapped main collection basic historic group has GPP and GGPP values that aredifferent from those for the corresponding non-main collection group, due to the fact that they have been updated historically using different previous-trip driving data sets. ln a step K, the extended previous-trip driving data set is collected at time point 102, that isthe same or corresponding observation time point as the above described previous-tripdriving data set. The extended driving data set is mapped, using an extended group con-formance measure, to a corresponding extended historic group, among a set of many suchextended historic groups. This mapping entirely corresponds to the mapping to the basichistoric groups, as described above, and is also based upon identical intervals. The mappedextended group is further mapped to the main collection basic historic group describedabove, using the knowledge available to the system that the basic and extended previous- trip driving data sets, respectively, were collected at the same time (at time point 102).

Then, in a step N, a GPP value of the mapped extended group is updated using the GPP value of said mapped main collection basic historic group, using an average function corre- 67 sponding to the ones described above. Hence, the GPP value of the extended group is up-dated, based upon the GPP value 102.89 of the basic group, from 101.42 to 101.44, reflect-ing the fact that the GPP value of 102.89 is higher than 101.42. ln the case the previous trip is seen as a current trip, the same steps A-J are performed, withthe goal of not only updating the basic historic group data, but also to calculate said first and second trip performance parameter values for presentation to the current driver. ln the particular case in which the current vehicle does not offer fuel consumption data, theupdates in steps G, I andJ are not performed, since there is no data available for doing thoseupdates. However, the first and second trip performance parameters may still be calculated and presented to the current driver. ln case the previous or current vehicle is not a trusted vehicle, the updates are performedbut a respective quality flag on each historic basic group may not be set to indicate full quality. ln the particular case in which no basic parameter data is available for readout from thecurrent vehicle, steps A-J are not performed at all. lnstead, steps K, L and M are performed,and the first and second trip performance parameters are calculated based upon themapped extended groups corresponding to each collected extended previous-trip driving data set. The update in step N is not performed in this case.

Hence, for a current vehicle without an interface providing data on fuel consumption, thesystem may calculate trip performance values for the current trip. Even for a current vehiclewithout any readable data whatsoever trip performance values can be calculated for thecurrent trip, as long as the extended data set is available for collection via a smartphone or other device present in the vehicle during the current trip.

Figure 17 illustrates the screen 132 of device 130 (or any other screen in the current vehicle or elsewhere which is accessible to the current driver during or after a current trip) at the 68 time of presenting the said information. On the screen 132, information (BM = benchmarkvalue, the historic 10% top performing trips with respect to first trip performance parame-ter; Efficiency = first trip performance parameter value for the current driver; and Risk =second trip performance parameter value for the current driver) provided from the centralserver 150 (or the local server 160, as the case may be) is displayed along a time axis (X-axis;the parameter values are displayed on the Y-axis), with one respective data point for eachof the last three trips performed by the driver in question, in this case regardless of whatvehicle was used. As can be seen in figure 17, the current driver has improved somewhatduring his or her last three current trips. The current driver has a slightly higher risk scorethan driving efficiency score, but still has some ways up to being among the best-performingdrivers. At the same time, the driver collective using the system has improved on average, increasing the BM value slightly over time. ln case the methodology described herein is followed for calculating the said first and sec-ond trip performance parameters, it has turned out that the first trip performance param-eter is an accurate measure of relative driving efficiency, regardless of vehicle, and that thesecond trip performance parameter is an accurate measure ofdriving riskiness, also regard- less of vehicle.

Above, preferred embodiments have been described. However, it is apparent to the skilledperson that many modifications can be made to the disclosed embodiments without de- parting from the basic idea ofthe invention. lt is generally noted, that the above described seven aspects of the present invention arefreely combinable in any constellation, and individual details from one of said aspects are readily useful in any ofthe other aspects, as applicable. ln general, it is preferred that the present system does not perform any analysis ofthe cur-rent trip based upon geographic location of the vehicle or based upon map data. lnstead,the system preferably completely relies upon current-trip driving data sets as collected dur- ing each trip and as compared to previous-trip driving data sets as described above. 69 lf the said basic historic groups are used, it is realized that the actual previous-trip drivingdata sets need not be stored in the database at all. lnstead, after a previous-trip driving dataset has been mapped to a basic historic group, and the basic historic group has been up-dated, such as with respect to its group performance parameter, the previous-trip drivingdata set may actually be discarded and not stored. Then, the information comprised in theprevious-trip driving data set lives on in the database in the form of the definition of the basic historic group in combination with the updated group performance parameter value. l\/loreover, each driving data set parameter value may be an instantaneously read value, orbe measured over a certain small time period, such as about 1 second, and averaged across that small time period. lt is realized that vehicles of fundamentally different types, such as gasoline vehicles, com-pletely electrical cars, boats, aeroplanes and bicycles, are preferably allotted to differentinstantiations of the present system, in order to achieve more relevant data comparisonsbetween different vehicle classes. However, it is also possible to one single system for allsuch different vehicle types, since the vehicle classes will typically converge into a set ofclasses wherein different types of vehicles are properly represented, as long as the basicparameter set, the qualified parameter set and the extended parameter set (as applicable) are ca refully selected.

Hence, the invention is not limited to the described embodiments, but can be varied within the scope ofthe enclosed claims.

Common expressions and definitions Current trip = the trip which is performed by a particular driver and a particular vehiclenow, and for which a trip performance parameter value is to be calculated.

Previous trip = a trip which was performed at least partially before the current trip.Predetermined set of basic driving data parameters = basic parameter set = standarddata set provided by vehicle.

Basic data set = observed set of parameter data comprising the basic parameter set.Predetermined set of qualified driving data parameters = qualified parameter set = basicparameter set as well as instantaneous energy consumption.

Qualified data set = observed set of parameter data comprising the qualified parameterset.

Predetermined set of extended driving data parameters = extended parameter set =standard data set not entirely provided by vehicle.

Extended data set = observed set of parameter data comprising the extended parameterset.

Current-trip driving data set = set of parameter data observed during the current trip. Acurrent-trip data set can be a current basic data set, a current qualified data set and/or acurrent extended data set.

Previous-trip driving data set = set of parameter data observed during a previous trip. Aprevious-trip data set can be a previous basic data set, a previous qualified data set and/ora previous extended data set.

Historic basic group of previous-trip driving data sets = basic historic group = the previ-ous-trip data sets that are ”similar” to each other according to the basic similarity meas-ure.

Historic extended group of previous-trip driving data sets = extended historic group =the previous-trip driving data sets that are ”similar” to each other according to the ex-tended similarity measure.

Basic driving data set similarity measure = basic similarity measure = comparison meas- ure for basic and qualified parameter sets. 71 Extended driving data set similarity measure = extended similarity measure = compari-son measure for extended parameter sets.

Basic conformity measure for a driving data set to a historic group of previous-trip driv-ing data sets = Basic group conformity measure = conformity measure between a basic orqualified parameter set and a basic historic group.

Basic conformity measure for the driving data sets for a particular vehicle to a set ofclass-defining parameters = Basic class conformity measure = conformity measure be-tween a number of driving data sets for a vehicle and a certain parameterized characteris-tic information for a particular class of vehicles.

Extended conformity measure for a current-trip driving data set to a historic group ofprevious-trip driving data sets = Extended group conformity measure = conformity meas-ure between an extended parameter set and an extended historic group.

Collection of previous-trip data sets = collection of previous-trip driving data sets or basichistoric groups for all vehicles belonging to a certain class of vehicles.

Current class = the class to which the current vehicle belongs.

Current collection = the collection corresponding the class to which the current vehiclebelongs.

Energy consumption-based group performance parameter = Energy-based group perfor-mance parameter = performance parameter calculated based upon energy consumptionfor previous-trip driving data sets of a particular basic historic group.

General group performance parameter = General group performance parameter = per-formance parameter for a particular basic historic group calculated based upon respectivevalues of energy -based group performance parameters for other basic historic groups.First energy consumption-based trip performance parameter = First energy-based tripperformance parameter = performance parameter calculated for a current trip basedupon energy consumption-based group parameters for basic historic groups.

Second energy consumption-based trip performance parameter = Second energy-basedtrip performance parameter = performance parameter calculated for a current trip basedupon first energy-based trip performance parameter values for basic historic groups.Current vehicle = the vehicle currently being driven, for which the performance parameter is to be calculated 72 Driver = person or entity driving or controlling vehicleCharacteristic instantaneous relative energy consumption curve = Function describing, for a particular vehicle class, a typical relationship between instantaneous vehicle velocity and relative energy consumption-based performance.

Claims (12)

1. Method for automatically assessing performance of a driver (110) of a current vehicle (100) for a particular current trip, wherein updated current-trip driving data sets are repeat- 5 edly read from the vehicle (100), which current-trip driving data sets each comprises data from at least a predetermined set of basic driving data parameters, wherein new such cur- rent-trip driving data sets are read from the vehicle (100) at consecutive observation time points separated by at the most a predetermined observation time period, c h a r a c - terised 10 a) b)20 c) d) i n that the method comprises the steps of collecting previous-trip driving data sets, observed at a plurality of different observa-tion time points, for a plurality of different previous trips made by a plurality of differ-ent drivers and a plurality of different vehicles, which previous-trip driving data setseach comprises parameter values for at least a certain predetermined set of qualifieddriving data parameters in turn comprising the said basic parameter set and in partic-ular instantaneous vehicle energy consumption and instantaneous vehicle velocity;for a plurality of said previous-trip driving data sets, calculating a respective relativeinstantaneous vehicle energy consumption value, which relative energy consumptionis relative to a total energy consumption for a respective trip during which the previ-ous-trip driving data set in question was observed; calculating a characteristic vehicle relative energy consumption function regardingthe value of said relative instantaneous vehicle energy consumption for different in-stantaneous vehicle velocity parameter values; and calculating a value of a trip performance parameter based upon a weighted averagevalue of the respective relative instantaneous energy consumptions for previous-tripdriving data sets that correspond to each of said current-trip driving data sets basedupon a similarity or conformance measure regarding the respective values of saidbasic parameters, which weighting is performed using said characteristic vehicle rela- tive energy consumption function. 74

2. Method according to claim 1, c h a r a c t e r i s e d i n that said character-istic vehicle relative energy consumption function is calculated based upon an average rel-ative instantaneous vehicle energy consumption for several previous-trip driving data setshaving the same vehicle velocity.

3. Method according to claim 1 or 2, c h a r a c t e r i s e d i n that said pre-vious-trip driving data sets are classified into one of a plurality of different predeterminedbasic historic groups based upon a basic similarity measure arranged to measure similarity between driving data sets, and in that each current-trip driving data set is mapped to at the most one of said basic historic groups based upon said basic group conformity measure.

4. Method according to claim 3, c h a r a c t e r i s e d i n that the said char-acteristic vehicle relative energy consumption function is calculated based upon a pluralityof such groups, and specifically upon a respective value of said relative instantaneous vehi-cle energy consumption for the previous-trip driving data sets belonging to the respectivebasic historic group in question.

5. Method according to claim 3 or 4, c h a r a c t e r i s e d i n that said rela-tive instantaneous vehicle energy consumption value is calculated for the respective previ-ous-trip driving data sets in the basic historic group to which the current-trip driving data set in question is mapped, in relation to a total energy consumption for the complete trip during which the previous-trip driving data set in question was observed.

6. Method according to any one of the preceding claims, c h a r a c t e r i s e d i n that the method further comprises the step ofclassifying said previous-trip driving datasets into a set of collections, wherein each of said collections only comprises previous-tripdriving data sets for a particular class of vehicles, wherein the current vehicle is classifiedinto a particular current class ofa set ofclasses based upon a basic class conformity measurebetween driving data sets for the vehicle in question and a set of class-defining parameters,said class corresponding to a current collection of previous-trip driving data sets, whereinall previous-trip driving data sets of one and the same vehicle (100) are classified into one and the same collection, based upon said basic class conformity measure; wherein the said trip performance parameter value is calculated based upon only the said respective relativeinstantaneous vehicle energy consumption values for previous-trip driving data sets in said current collection.

7. Method according to claim 6, c h a r a c t e r i s e d i n that the said basicparameter set comprises instantaneous vehicle velocity and instantaneous engine rotationspeed, and wherein the said class-defining parameters comprise, for each class of vehicles, a characteristic engine rotation speed for a particular vehicle velocity.

8. Method according to any one of the preceding claims, c h a r a c t e r i s e di n that said predetermined observation time period is at the most 10 seconds, preferably at the most 5 seconds, more preferably at the most 2 seconds.

9. Method according to any one ofthe preceding claims, c h a r a c t e r i s e d i n that parameter values of said basic parameter set are automatically recorded by thevehicle (100) and either communicated to a portable electronic device (130) arranged atthe vehicle (100), which portable electronic device (130) communicates, via a wireless link,said parameter values to a central server (150), or communicated, via a wireless link, di- rectly from the vehicle (100) to said central server (150).

10. Method according to any one ofthe preceding claims, c h a r a c t e r i s e di n that the said trip performance parameter value is calculated by and communicated,via a wireless link, from a central server (150) to the vehicle (100), such as to a portable electronic device (130) arranged at the vehicle (100), and presented to the driver (110).

11. Method according claim 10, c h a r a c t e r i s e d i n that a value of saidtrip performance parameter is calculated repeatedly, preferably at least every 10 minutes,more preferably at least every 2 minutes, even more preferably at least every 30 seconds,during the current trip, wherein the current trip is considered to be the current trip up tothe moment at which the value ofthe said trip performance parameter is calculated and for the purposes of calculating the said trip performance parameter value in question, and in 76 that the currently calculated such value is communicated to the vehicle (100) and presented to the driver (110) upon calculation.

12. System for automatically assessing performance of a driver (110) of a current vehicle(100) for a particular current trip, which system is arranged to repeatedly read updatedcurrent-trip driving data sets from the vehicle (100), which current-trip driving data setseach comprises data from at least a predetermined set of basic driving data parameters,wherein the system is arranged to read new such current-trip driving data sets from thevehicle (100) at consecutive observation time points separated by at the most a predeter-mined observation time period, c h a r a c t e r i s e d i n that the system com-prises a server (150,160), arranged to collect previous-trip driving data sets, observed at aplurality of different observation time points, for a plurality of different previous trips madeby a plurality of different drivers and a plurality of different vehicles, which previous-tripdriving data sets each comprises parameter values for at least a certain predetermined setof qualified driving data parameters in turn comprising the said basic parameter set and inparticular instantaneous vehicle energy consumption and instantaneous vehicle velocity, inthat the server (150,160) is arra nged to, for a plurality of said previous-trip driving data sets,calculate a respective relative instantaneous vehicle energy consumption value, which rel-ative energy consumption is relative to a total energy consumption for a respective tripduring which the previous-trip driving data set in question was observed, in that the server(150,160) is arra nged to calculate a cha racteristic vehicle relative energy consumption func-tion regarding the value of said relative instantaneous vehicle energy consumption for dif-ferent instantaneous vehicle velocity parameter values, and in that the server (150,160) isarranged to calculate a value of a trip performance parameter based upon a weighted av-erage value of the respective relative instantaneous energy consumptions for previous-tripdriving data sets that correspond to each of said current-trip driving data sets based upona similarity or conformance measure regarding the respective values of said basic parame-ters, which weighting is performed using said characteristic vehicle relative energy con- sumption function.